This file is the central point of navigation, from which you can jump to a specific session or paper by clicking on the title in the Paper Index. Having done so, information about the paper (abstract, authors, etc.) will be shown, with the option of downloading the full paper.

You can also press CTRL+F to search through this file, helping you find a desired paper based on the search keywords you enter, which can be of the topic of the paper, its title, name of the authors, etc.

Tu01. Control Theory: Multivariable systems and networks
   Tu01.01: Nonlinear pole assignment control of state dependent parameter models with time delays
   Tu01.02: A multiple-observer approach to stability in wireless network control systems
   Tu01.03: Hinf-based model order reduction using LMIs
   Tu01.04: The Investigation of Multivariable Control Performance Assessment Techniques
   Tu01.05: Pole Placement Controller Design for Linear Parameter Varying Plants
   Tu01.06: ROBUST $H_\infty$ CONTROL FOR NEUTRAL SYSTEMS VIA DYNAMIC OUTPUT FEEDBACK

Tu02. Control of nonlinear systems
   Tu02.01: Output Feedback Sampled-Data Control of Nonlinear Systems in Output Feedback Form
   Tu02.02: Constraint handling for State Dependent Parameter models
   Tu02.03: Robust Adaptive Nonlinear Control Law for a General Class of Nonlinear Systems with Operator-Based Hysteresis Models
   Tu02.04: Dynamical Radial Control of Nonlinear Systems
   Tu02.05: CONTROL OF POLYETHYLENE PROPERTIES USING NONLINEAR MODEL PREDICTIVE CONTROL
   Tu02.06: Asymptotic Rejection of Nonlinear Periodic Disturbances in Linear Dynamic Systems

Tu03. Human Adaptive Mechatronics (Invited)
   Tu03.01: Dynamic Model of Muscle Force Driving System and Its Application in Tele-operation
   Tu03.02: Control of a propulsion mechanism over a wireless network
   Tu03.03: An Iterative Learning Control Scheme for the Capsubot
   Tu03.04: Combined Attitude Control of an Underactuated Helicopter Experimental System
   Tu03.05: Complex Motor Cortex Control of Muscle Synergies Underpin Simple Reaching Tasks in Robot-Induced Force Fields
   Tu03.06: Insights into Information Processing by the Single Cell Slime Mold Physarum Polycephalum0

Tu04. Estimation and filtering of nonlinear and delayed systems
   Tu04.01: Identification Applied to Dual Sensor Transient Temperature Measurement
   Tu04.02: The implementation of simulated annealing combining gradient search in system identification
   Tu04.03: Design and Real Time Implementation of Nonlinear Minimum Variance Filter
   Tu04.05: Constrained particle filtering using Gaussian sum approximations
   Tu04.06: Time-delay in high-gain observer based disturbance estimation

Tu05. Fault detection and plant monitoring
   Tu05.01: Discrete-time Robust Fault Detection Observer Design: a simulated annealing approach
   Tu05.02: SENSOR LOCATION BASED OPTIMUM DESIGN FOR FAULT DETECTION SYSTEM
   Tu05.03: Model-Based Sensor Fault Diagnosis in General Stochastic Systems Using LMI Techniques
   Tu05.04: Multi-Agent Control of High Redundancy Actuation
   Tu05.05: Detection of Additive Sensor Faults in an Unmanned Air Vehicle (UAV) Model using Neural Networks
   Tu05.06: statistical process monitoring of bioreactors: a comparison

Tu06. Control Theory: Optimization and design
   Tu06.01: A Design-Orientated Approach to the Geometry of Fundamental Design Limitations
   Tu06.02: 'Flat Phase' PID Controllers
   Tu06.03: A reduced structure controller for a Grinder Circuit system
   Tu06.04: Comparison between MSE and MEE Based Component Extraction Approaches to Process Monitoring and Fault Diagnosis
   Tu06.05: Using a Chebyshev approach for the minimum-time open-loop control of constrained MIMO systems
   Tu06.06: PROPERTIES OF OUTPUT FREQUENCIES OF VOLTERRA SYSTEMS

Tu07. Mechatronics
   Tu07.01: Fault Detection Using High Gain Observer: Application in Pipeline System
   Tu07.02: Control of a simple DC motor robot equipped with ultrasonic sensors via a field programmable gate array and a speech recognition board and microphone
   Tu07.03: Heuristiscs-based High-level Strategy in Multi-Agent Environment
   Tu07.04: STRUCTURAL ANALYSIS OF THE DEFORMATION OF THE FLEXIBLE ARM OF A ROBOT: PART 2

Tu08. Adaptive Control
   Tu08.01: LINEARISATION OF POWER AMPLIFIERS, USING MINIMAL CONTROL SYNTHESIS
   Tu08.02: A METHOD FOR FINDING GOOD VALUES OF ADAPTATION GAINS
   Tu08.03: U-MODEL BASED ADAPTIVE INTERNAL MODEL CONTROL OF UNKNOWN MIMO NONLINEAR SYSTEMS: A CASE STUDY ON 2-LINK ROBOTIC ARM
   Tu08.04: A Disturbance Rejection Supervisor in Multiple-Model Based Control
   Tu08.05: Automatic Learning in Multiple Model Adaptive Control
   Tu08.06: EXPERIMENTAL IMPLEMENTATION AND VALIDATION OF DUAL ADAPTIVE CONTROL FOR MOBILE ROBOTS

Tu09. ACE: Invited session
   Tu09.01: CONTROL OF INTEGRAL PROCESSES WITH DEAD TIME: PRACTICAL ISSUES AND EXPERIMENTAL RESULTS
   Tu09.02: An Assessment of a Modified Optimal Control Strategy as Applied to the Control of an Unmanned Surface Vehicle
   Tu09.03: A Fast Training Algorithm For Least-Squares Support Vector Machines
   Tu09.04: INTELLIGENT SOFTWARE SENSORS FOR FED-BATCH FERMENTATION PROCESSES
   Tu09.05: An economic parametrisation for parahermitian matrix functions used in control systems optimisation
   Tu09.06: Robust Control of a High Redundancy Actuator

Tu10. Predictive Control
   Tu10.01: A Model Predictive Approach to Wireless Networked Control
   Tu10.02: Design of Reconfigurable Predictive Control Applied on the Air Path of a Diesel Engine
   Tu10.03: Model Predictive Control of Substructured Systems
   Tu10.04: Constrained Predictive Control Of A Servo-driven Tracking Turret
   Tu10.05: AQM Control of TCP/IP Networks using Generalized Predictive Control
   Tu10.06: Subspace-based Model Predictive Control with Data Prefiltering

Tu11. Control Theory: Optimization and nonlinear design
   Tu11.01: Controller Design of Conflict Multi-objective control problem by Preference
   Tu11.02: Control Engineering at the University of Manchester in the Post War Years
   Tu11.03: An Arnoldi Based Method to Discrete Time Linear Optimal Multi-periodic Repetitive Control
   Tu11.04: A non-parametric method for nonlinear J-J' spectral factorisation.
   Tu11.05: Design of Retarded Fractional Delay Differential Systems by the Method of Inequalities

Tu12. Decision and control (Invited)
   Tu12.01: Decision-Making of Football Agents with Support Vector Machine
   Tu12.02: OPERATOR BASED ROBUST NONLINEAR CONTROL SYSTEM DESIGN OF MIMO NONLINEAR FEEDBACK CONTROL SYSTEMS
   Tu12.03: LS-SVM based motion control of a mobile robot in dynamic environment
   Tu12.04: OPERATOR BASED FAULT DETECTION SYSTEM DESIGN TO AN ACTUATOR FAULT OF A THERMAL PROCESS
   Tu12.05: COMARISION OF CONTROL DESIGN TECHNIQUES FOR A NUCLEAR REACTOR

Tu13. Fuzzy Control Systems
   Tu13.01: A new engineering method for fuzzy reliability analysis of surge detection and isolation in centrifugal compressor
   Tu13.02: MODELLING AND CONTROL OF FES-ASSISTED INDOOR ROWING EXERCISE
   Tu13.03: Adaptive Fuzzy Model-based Predictive Control Using Fuzzy Decision Making
   Tu13.04: FUZZY CONTROLLER DEVELOPMENT FOR A PEM FUEL CELL SYSTEM
   Tu13.05: Reinforcement Learning for Probabilistic Fuzzy Controllers

Tu14. Systems control using neural networks
   Tu14.01: Controller Design for Nonlinear Systems with Stochastic Time Delays Using Neural Networks and Information Entropy
   Tu14.02: Controller Design of Nonlinear TITO Systems with Uncertain Delays via Neural Networks and Error Entropy Minimization
   Tu14.03: Single Network Adaptive Critic for Vibration Isolation Control
   Tu14.04: Discrete-Time Decentralized Neural Identification and Control for a 2 DOF Robot Manipulator
   Tu14.05: Neural Predictive Control for Wide Rage of Process Systems

Tu15. Modeling and Simulation
   Tu15.01: MODELLING, PARAMETER ESTIMATION AND VALIDATION OF A 300W PEM FUEL CELL SYSTEM
   Tu15.02: Modelling and Parameter Identification of Electrochemical Cu-Cu Cell
   Tu15.03: Phase Model for the relaxed van der Pol oscillator and its application to synchronization analysis
   Tu15.04: FPGA IMPLEMENTATION OF WHEEL-RAIL CONTACT LAWS

We01. Control Applications: Low complexity control
   We01.01: Study of Reduced-order and Non-linear Local Optimal Control Application to Aero Gas Turbines
   We01.02: Equalisation Tuning Method
   We01.03: Low Complexity Control of Hybrid Systems with Application to Control of Step-down DC-DC Converters
   We01.04: A Comparative study on charge system modelling in fine paper production
   We01.05: Robust H-Infinity Control of a Steerable Marine Radar Tracker
   We01.06: IDENTIFICATION OF PARKINSON’S DISEASE TREMOR ONSET USING ARTIFICIAL NEURAL NETWORKS

We02. Control Applications : Optimization and Networks
   We02.01: Nonparametric Collocation ODE Parameter Estimation: Application in Biochemical Pathway Modelling
   We02.02: Load Minimization Design for Internet-based Control
   We02.03: Adaptive Feedforward Control via Virtual Error Approach with Application to Predistortion of Nonlinear HPA
   We02.04: Design and Implementation of Brushless Motor Controller Based on SOPC
   We02.05: Motion stabilization in the presence of friction and backlash: a hybrid system approach

We03. Advanced Process Control
   We03.01: Modeling and Control of a Fluidised Bed Dryer
   We03.02: MEMBRANE MODELING FOR SIMULATION AND CONTROL OF REVERSE OSMOSIS IN DESALINATION PLANTS
   We03.03: CONTROLLING WATER QUALITY USING REVERSE OSMOSIS: THE DEVELOPMENT OF SIMPLIFIED DYNAMIC MODEL
   We03.04: Noncausal open-loop control with combined system identification and PID controller tuning

We04. Control Applications: Automotive
   We04.01: Fuel consumption optimization for a city bus
   We04.02: Semi-active ride control of human seated model and robustness analysis.
   We04.03: Asymptotic Tracking applied to the Control of a Turbocharged Diesel Engine
   We04.04: Constrained Variance Control of Peak Pressure Position by Spark Ionization Feedback

We05. Hybrid Vehicles (invited)
   We05.01: Modelling and Control of a novel SOFC-IC Engine Hybrid System
   We05.02: ELECTRICAL ARCHITECTURES FOR HYBRID VEHICLES: IMPLICATIONS FOR MODELLING AND CONTROL
   We05.03: A highly modular simulation model for hybrid electric fuel cell power drive trains
   We05.04: NEDC Based Compensated Forward Simulation Approach with Energy Management for Parallel Hybrid Electric Vehicles
   We05.05: Block-Control Methods for Low-Order Automotive Control

We06. Robotics: Vision and tracking
   We06.01: Visual Tracking System for the Welding of Narrow Butt Seams in Container Manufacture
   We06.02: Homing, Calibration and Model-Based Predictive Control for Planar Parallel Robots
   We06.03: Visual servoing control for line and object detection and following using a robotic arm manipulator mounted real time camera system
   We06.04: Walking Control Algorithm based on Polynomial Trajectory Generation
   We06.05: Experimental Evaluation of Haptic Control for Human Activated Command Devices
   We06.06: Path Planning Generation in Mobile Robots using Evolutionary Harmonic Potential Field Technique

We07. Control Applications : Aerospace
   We07.01: Robust, Power Aware Mobile Agent Tracking using an 802.15.4 Wireless Sensor Network.
   We07.02: Real-time trajectory generation technique for dynamic soaring UAVs
   We07.03: A Lateral Directional Flight Control System for the MOB Blended Wing Body Planform
   We07.04: Suppressing aeroelastic vibrations via stability region maximization and numerical continuation techniques
   We07.05: ANFIS Network Design Method for Modelling of the Twin Rotor MIMO System (TRMS)
   We07.06: The optimisation of stator vane settings in multi-stage axial compressors using a particle swarm optimisation

Th01. Control Theory: Discrete systems
   Th01.01: IMPLEMENTATION OF NON-UNIFORM SAMPLING FOR ‘ALIAS-FREE PROCESSING’ IN DIGITAL CONTROL
   Th01.02: Extraproximal Method for Markov Chains Finite Games
   Th01.03: $L_2 $ gain analysis for linear discrete switched delay systems
   Th01.04: Development of second order plus time delay (SOPTD) model from orthonormal basis filter (OBF) model
   Th01.05: Improved FOPDT model estimation with Delayed-relay feedback for constant time dominant processes
   Th01.06: Reduced-order Local Optimal Controller for a Higher Order System

Th02. Robotics: Control and recognition
   Th02.01: EXPERIMENTAL STUDIES OF MULTI-ROBOT FORMATION AND TRANSFORMING
   Th02.02: Control Laws Design and Simulation Validation of Autonomous Mobile Robot Off-Road Trajectory Tracking
   Th02.03: Fast Gabor Filters for Object Recognition of Mobile Robot
   Th02.04: Sub-Optimal Control Based on Passivity for Euler-Lagrange Systems.
   Th02.05: Control Based on Energy for Vertical 2 Link Underactuated Robots.

Th03. Control Methodology 1
   Th03.01: CHEAP COMPUTATION OF OPTIMAL REDUCED MODELS USING SYMBOLIC COMPUTATION
   Th03.02: On the Relative Degrees and the Interactor Matrix of Linear Multivariable Systems
   Th03.03: An Approach to Pole Placement Method with Output Feedback
   Th03.04: Robust output-feedback tracking control of multivariable continuous-time systems in an LMI setting
   Th03.05: Stabilizing systems with aperiodic sample-and-hold devices: state feedback case
   Th03.06: ROBUST CONTROLLER TUNING BASED ON COEFFICIENT DIAGRAM METHOD

Th04. Condition monitoring and fault diagnosis
   Th04.00: Fault Detection for Vehicle Suspensions Based on System Dynamic Interactions
   Th04.01: Observer-Based Residual Design for Nonlinear Systems
   Th04.02: Nonlinear PCA for Transient Monitoring of an Automotive Engine
   Th04.03: APPLICATION OF A PCA MODEL APPROACH FOR MISFIRE MONITORING
   Th04.04: Robust Fault Isolation for Autonomous Coordination in NCS

Th05. Parameter estimation and data analysis
   Th05.01: multivariate statistical analysis of spectroscopic data
   Th05.02: Novel algorithms based on conjunction of the Frisch scheme and extended compensated least squares
   Th05.03: Parameter Identification for Electromechanical Servo Systems Using a High-gain Observer
   Th05.04: Minimum Entropy Parameter Estimation of Bounded Nonlinear Dynamic Systems with Non-Gaussian state and Measurement noise
   Th05.05: Condition Monitoring Approaches to Estimating Wheel-Rail Profile
   Th05.06: Dynamic Model for the LHIfAM Haptic Interface: Friction parameter estimation

Th06. Control Theory: Uncertain and time varying
   Th06.01: Relay feedback based monitoring and autotuning of processes with gain nonlinearity
   Th06.02: Design and Implementation of a Time Varying Local Optimal Controller based on RLS Algorithm for Multivariable Systems
   Th06.03: CONTROL TECHNIQUES FOR MULTI-AXIS REAL-TIME DYNAMIC SUBSTRUCTURING
   Th06.04: FDI OF THREE-TANK SYSTEM USING NEUROFUZZY NETWORKS WITH LOCAL APPROACHES
   Th06.05: A foray into P2BL in a Control Systems Course
   Th06.06: A NEW APPROACH TO INPUT-OUTPUT PAIRING ANALYSIS FOR UNCERTAIN MULTIVARIABLE PLANTS

Th07. Inequality Procedures(invited)
   Th07.01: Method of Inequality-Based Multiobjective Genetic Algorithm for Optimizing Cart-Double-Pendulum-System
   Th07.02: Design of Critical Control Systems Using Disturbance Cancellation Controllers
   Th07.03: Development of the actively-controlled beds for ambulances
   Th07.04: Robust Multivariable Control System Design Using The Method Of Inequalities
   Th07.05: Poiseuille Flow Controller Design via the Method of Inequalities
   Th07.06: Computation of Peak Output for Inputs Restricted in $\mathcal{L}_2$ and $\mathcal{L}_\infty$ Norms Using Convex Optimization

Th08. Sliding Mode Control
   Th08.01: SLIDING MODE CONTROLLERS USING OUTPUT INFORMATION: AN LMI APPROACH
   Th08.02: Application of MPC and Sliding Mode Control To IFAC Benchmark Models
   Th08.03: SLIDING-MODE POSITION CONTROL OF A 1-DOF SET-UP BASED ON PNEUMATIC MUSCLES
   Th08.04: About Equivalence Between Sliding Mode and Continuous Control Systems
   Th08.05: Fuzzy Sliding Mode Controllers for Vehicle Active Suspensions
   Th08.06: Design of an Asymptotic Sliding Mode Algorithm for Nonlinear Systems: An Observer Based Approach

Th09. Imaging and Road Traffic Control
   Th09.01: Movement-Based Look-Ahead Traffic-Adaptive Intersection Control
   Th09.02: Development of Knowledge-based Measurement with Monocular Vision
   Th09.03: Multiple Kernel Learning from Sets of Partially Matching Image Features
   Th09.04: A STUDY ON THE EFFECT OF GPS ACCURACY ON A GPS/INS KALMAN FILTER
   Th09.05: Control Schemes for Safe Operation of Vehicles Convoys

Th10. Control Methodology 2
   Th10.00: A hands-on approach toward vehicle velocity estimation
   Th10.01: DATA-DRIVEN DIRECT ADAPTIVE MODEL BASED PREDICTIVE
   Th10.02: A New Multi Agent Approach for Traffic Shaping and Buffer Allocation in Routers
   Th10.03: Using Lagged Spectral Data in Feedback Control Using Particle Swarm Optimisation
   Th10.04: Exact Controls for Superconformal Via Fill Process

Tu01. Control Theory: Multivariable systems and networks

Tu01.01: Nonlinear pole assignment control of state dependent parameter models with time delays
Click here for full paper
Authors: Dr. C. James Taylor     Dr. Arun Chotai     Prof. Peter C. Young    
Abstract: This paper considers pole assignment control of nonlinear dynamic systems described by State Dependent Parameter (SDP) models. The approach follows from earlier research into linear Proportional-Integral-Plus (PIP) methods but, in SDP system control, the control coefficients are updated at each sampling instant on the basis of the latest SDP relationships. Alternatively, algebraic solutions can be derived off-line to yield a practically useful control algorithm that is relatively straightforward to implement on a digital computer, requiring only the storage of lagged system variables, coupled with straightforward arithmetic expressions in the control software. Although the analysis is limited to the case when the open-loop system has no zeros, time delays are handled automatically. The paper shows that the closed-loop system reduces to a linear transfer function with the specified (design) poles. Hence, assuming pole assignability at each sample, global stability of the nonlinear system is guaranteed at the design stage. The associated conditions for pole assignability are stated.
Return to index

Tu01.02: A multiple-observer approach to stability in wireless network control systems
Click here for full paper
Authors: Mr Adrian McKernan     Dr Carlos Ariño     Prof. George Irwin     Dr William Scanlan     Mr Jian Chen    
Abstract: This paper describes a new multiple-observer approach to Wireless Network Control Systems (WNCS). Two sets of observers are proposed, Lost Sample Observers (LSO) to deal with packet dropout and State Prediction Observers (SPO) to compensate for time-varying delays. These are designed using Linear Matrix Inequalities (LMI), thereby ensuring closed-loop stability. A numerical example, of a cart-mounted inverted pendulum is given along with results from simulation studies, comparing this new approach with a constant gain Linear Quadratic Regulator (LQR), in the presence of time-varying delays. Practical experimental results, on a IEEE 802.11b wireless channel in a reverberation chamber, further confirm the efficacy of the approach.
Return to index

Tu01.03: Hinf-based model order reduction using LMIs
Click here for full paper
Authors: Dr Amin Nobakhti     Prof Hong Wang    
Abstract: In this paper new sufficient conditions are presented for the existence of a Lyapunov pair with a coupling rank constraint within a $\mathcal{H}_{\infty}$ minimization framework derived using the bounded real lemma and the projection lemma. The conditions are then used to propose a Linear Matrix Inequality (LMI) sub-optimal method to solve the model order reduction problem of general non-square LTI systems with a prescribed number of states to be removed. This alleviates the need for trace or rank minimization, iterations, or a priori choice of any new additional variable. The effectiveness and stability of the proposed LMI method is demonstrated by applications to several model order reduction problems.
Return to index

Tu01.04: The Investigation of Multivariable Control Performance Assessment Techniques
Click here for full paper
Authors: Miss Qiaolin Yuan     Prof. Barry Lennox    
Abstract: Much attention has been paid to Control Performance Assessment (CPA) since the Harris Index was first proposed. This paper argues that there are two fundamental requirements for any CPA algorithm. The first is that it should be able to detect any change in the performance of a control system and the second is that it should be able to identify the potential improvement that can be made to the performance of the control system by re-tuning or re-configuring it. The ability of current multivariable CPA techniques to address these two issues is investigated in this paper, and limitations with the currently available approaches are identified and concluded in brief. The benefits of addressing the two issues are demonstrated using a simulated multivariate system, and the results of a detailed study identify a CPA approach which is able to address both of these issues and also diagnose the root cause of any change in control performance.
Return to index

Tu01.05: Pole Placement Controller Design for Linear Parameter Varying Plants
Click here for full paper
Authors: Mr Sunan Chumalee     Dr James Whidborne    
Abstract: Nonlinear plants can often be modelled as linear parameter varying (LPV) plants for which a number of techniques exist for control synthesis. However, there are some systems for which such a technique presents difficulties. In this paper, we consider one such system for which we propose a pole placement method using state feedback in order to cancel parameters variation of LPV plants. Hence, any linear time invariant (LTI) controller can subsequently be employed for an outer loop. The approach is demonstrated on an example for which only a single output can be measured. Therefore, a state observer for an LPV plant is also demonstrated in order to estimate state values. The simulation results show that the new approach yields reliable closed-loop stability with good closed-loop transient performance of the system.
Return to index

Tu01.06: ROBUST $H_\infty$ CONTROL FOR NEUTRAL SYSTEMS VIA DYNAMIC OUTPUT FEEDBACK
Click here for full paper
Authors: Dr. Bayram Baris Kizilsac     Prof. Ulviye Baser    
Abstract: This paper deals with the design problem for dynamic output feedback robust $H_\infty$ control problem for a class of uncertain linear neutral systems in delay dependent case. Sufficient conditions for the existence of controller is derived based on the delay dependent Bounded Real Lemma (BRL) in terms of linear matrix inequalities (LMIs) with inverse constraints which is obtained without resorting to any model transformations. A convex optimization algorithm is used to satisfy these constraints. A numerical example is given to illustrate the effectiveness of the proposed results.
Return to index

Tu02. Control of nonlinear systems

Tu02.01: Output Feedback Sampled-Data Control of Nonlinear Systems in Output Feedback Form
Click here for full paper
Authors: Dr. Zhengtao Ding     Mr. Buzhou Wu    
Abstract: This paper deals with sampled-data control of nonlinear systems in the output feedback form. A sampled-data control strategy is proposed based on the existing control design in the continuous-time domain via output feedback. The proposed control uses the sampled output and a discrete-time implementation of filters involved. The overall stability of the system under the proposed control has been analyzed, and the semi-global asymptotic stability for the system with relative degree one and two is established by keeping the sampling interval with a specified range.
Return to index

Tu02.02: Constraint handling for State Dependent Parameter models
Click here for full paper
Authors: Dr. Vasileios Exadaktylos     Dr. C. James Taylor     Dr. Arun Chotai    
Abstract: This paper considers Proportional-Integral-Plus (PIP) control of nonlinear dynamic systems described by State Dependent Parameter (SDP) models with constraints. More specifically, a low level stabilising SDP/PIP controller is first developed to steer the system in the desired direction, whilst a Reference Governor (RG) is subsequently introduced to account for constraints in the system variables. This contrasts with the (off-line) simulation-based methods previously used for PIP control of SDP models with constraints. Furthermore, the particular parametrisation of the RG used in this paper provides useful insight into, and quantification of, the effects of the constraints on the nonlinear control system.
Return to index

Tu02.03: Robust Adaptive Nonlinear Control Law for a General Class of Nonlinear Systems with Operator-Based Hysteresis Models
Click here for full paper
Authors: Prof. Chun-Yi Su     Dr. Qingqing Wang     Dr. Ying Feng     Prof. Shuzhi Ge    
Abstract: For the nonlinear systems preceded by smart actuators which exhibit hysteresis nonlinearities, it is a challenge to mitigate effects of the hysteresis. By utilizing an operator-based Prandtl-Ishlinskii model and a neural network approximator, a robust adaptive control scheme is developed for a general class of continuous-time nonlinear dynamic systems with unknown hysteresis nonlinearities. The boundedness of the closed-loop system is achieved and the tracking error converges to a set of adjustable neighborhood of zero independent of initial conditions. The effectiveness of the proposed control approach is demonstrated through a simulation example.
Return to index

Tu02.04: Dynamical Radial Control of Nonlinear Systems
Click here for full paper
Authors: Dr Zahra Sangelaji    
Abstract: This paper is concerned with the stabilisation of a general class of nonlinear systems via the associated angular approach. In this method, the system is converted into two subsystems the so called radial and spherical systems. The spherical system is a nonlinear equation on a sphere and the radial system is a scalar differential equation. A stabilising control can be designed based on the one-dimensional radial system dynamics. The radial control may be continuous or discontinuous depending on the structure of the input map. Whenever the input map of the radial subsystem is zero, the radial control is not accessible. In this paper a method is presented to remove this obstacle. The control is designed by including an extra dynamic to the system. Therefore the new system is an augmented system. The radial auxiliary input map of the augmented system i.e. the original control is the new state. Since it is assumed that the original control is not zero, the auxiliary radial control is definable within the operating region.
Return to index

Tu02.05: CONTROL OF POLYETHYLENE PROPERTIES USING NONLINEAR MODEL PREDICTIVE CONTROL
Click here for full paper
Authors: Dr. Mohammad Al-haj Ali     Prof. Emad Ali    
Abstract: This paper deals with the control of the melt index and density of polymers produced in gas phase polyethylene reactors. Nonlinear Model predictive control (NLMPC) is used for this purpose. A nonlinear reactor model combined with correlations for predicting polymer melt index and density are used to simulate the process. The simulations revealed the effectiveness of NLMPC to drive the polymer properties to follow a series of grade changeover in the absences and presence of modeling errors. Grade transition is achieved with zero offset but with relatively large settling time. Rapid grade changeover is limited by the large residence time and broad residence time distributions for both the gas phase and formed polymers.
Return to index

Tu02.06: Asymptotic Rejection of Nonlinear Periodic Disturbances in Linear Dynamic Systems
Click here for full paper
Authors: Dr Zhengtao Ding    
Abstract: This paper deals with asymptotic rejection of disturbances generated from nonlinear exosystems. The dynamic system is assumed to be linear. A new strategy for internal model design is proposed, based on a dynamic extension of the existing nonlinear observer design for the nonlinear exosystem. Additional filters are used to estimate the invariant manifold in the state space subject to the nonlinear disturbances generated from the exosystem. The proposed design for the internal model and control ensures that the state variable asymptotically converge to the invariant manifold, which implies that the designated output state asymptotically converge to zero.
Return to index

Tu03. Human Adaptive Mechatronics (Invited)

Tu03.01: Dynamic Model of Muscle Force Driving System and Its Application in Tele-operation
Click here for full paper
Authors: 1 Fu Xiuhui     2 Li Hongyi     3 Wang Yuechao    
Abstract: Muscle force model with constant degree of nervousness was proposed, in the context of modeling the operator system and operation delay of the internet-based teleoperation system. The dynamic model of the operator’s arm–joystick with force reflection was obtained. Dynamic compensation of muscle force driving system was given and verified by teleoperation experiments of a mobile robot through internet.
Return to index

Tu03.02: Control of a propulsion mechanism over a wireless network
Click here for full paper
Authors: Mr Sam Wane     Prof Hongnian Yu    
Abstract: This paper investigates the implementation of a pendulum-driven cart-pole system through wireless networks. The system is underactuated since the only control input is the motor which drives the pendulum movement while the cart has free movement wheels. An onboard client PC controls the torque to the motor, whilst a host PC monitors progress and controls the demand to the motor. The two PCs have been connected via a proprietary wireless network to allow the controller to be remote from the robot. The client PC interprets commands sent via the network from the host PC to control the torque of the pendulum device. The client PC also relays the pendulum position which the host PC uses as feedback to specify the torque to send.
Return to index

Tu03.03: An Iterative Learning Control Scheme for the Capsubot
Click here for full paper
Authors: Mr Yang Liu     Prof Hongnian Yu     Prof Luige Vladareanu    
Abstract: A Capsubot, which consists of three parts, the inner body, the capsule shell, and the driving mechanism, is a micro capsule robot with no legs and no wheels, and is driven by the interactive propulsion between the inner body and the capsule shell. The desired locomotion of the Capsubot is generated by making the inner body track a designed trajectory repeatedly. Due to the nature of repetitive motion, an iterative learning control scheme is proposed to improve the tracking performance of the inner body, in order to achieve the desired locomotion of the Capsubot. Extensive simulation studies are conducted to demonstrate the effectiveness of the scheme.
Return to index

Tu03.04: Combined Attitude Control of an Underactuated Helicopter Experimental System
Click here for full paper
Authors: Prof Mingcong Deng     Prof Akira Inoue     Mr Tatsunori Shimizu    
Abstract: In this paper, combined attitude control of an underactuated helicopter experimental system is considered. The controlled helicopter experimental system has two inputs and three outputs, namely, this system is underactuated. The combined attitude controller includes a nonlinear MIMO controller based on adaptive sliding mode control and non-adaptive nonlinear controllers. Control system stability is guaranteed by Lyapunov function based proof. Comparing simulation between the existed design method and the proposed design method shows the effectiveness of the proposed method.
Return to index

Tu03.05: Complex Motor Cortex Control of Muscle Synergies Underpin Simple Reaching Tasks in Robot-Induced Force Fields
Click here for full paper
Authors: Prof. Duncan Turner     Dr Paul Sacco     Mr Tim Hunter    
Abstract: In order to design effective human-machine interfaces, it is important to demonstrate that stereotypical movements such as reaching display predictable patterns of activation in muscles that operate at shoulder, elbow and wrist joints. Whilst humans display a wide repertoire of adaptive behavior in natural movements, this study demonstrates that muscles acting at different arm joints operate in synergies during reaching movements in a direction-dependent manner. These basic synergies can be mapped to similar direction-dependent motor cortex excitability maps and this plasticity of muscle and central nervous system should be taken into account when developing actuator systems which mimic natural movement.
Return to index

Tu03.06: Insights into Information Processing by the Single Cell Slime Mold Physarum Polycephalum0
Click here for full paper
Authors: Dr Steve Hickey     Dr Len Noriega    
Abstract: The finding that a simple single-celled organism can traverse a maze near optimally provides a challenge to some current ideas in artificial intelligence. In this paper, we present a simple explanation for such behavior and a computational model based on decision trees and ant algorithmics. The behavior of simple biological organisms may provide insights into the nature and evolution of intelligence.
Return to index

Tu04. Estimation and filtering of nonlinear and delayed systems

Tu04.01: Identification Applied to Dual Sensor Transient Temperature Measurement
Click here for full paper
Authors: Mr Colin Brown     Prof. George Irwin     Dr Robert Kee     Dr Sean McLoone     Dr Peter Hung    
Abstract: The harsh environment presented by engines, particularly in exhaust systems,necessitates the use of robust and therefore low bandwidth temperature sensors. Consequently,high frequencies are attenuated in the sensor output. A number of techniques for addressing this problem involve measurement of the gas temperature using two thermocouples with different time-constants and mathematical reconstruction of the true gas temperature from the resulting signals. Many of these methods rely on the assumption that the ratio of the thermocouple time-constants is invariant and known a priori. In addition, they are generally subject to singularities and sensitive to noise. A recently proposed two-thermocouple sensor characterization method which utilises system identification techniques and is much more generally applicable is described. Previous offline methods for constant velocity flow are extended using polynomial parameter fitting on a sliding data window to accommodate variable velocity. These methods have been successfully tested and proven for the first time in variable velocity flow with experimental data produced from a novel and highly instrumented test rig. Results show that the increase in bandwidth arising from the dual sensor technique allowed accurate measurement of fluctuating temperatures with relatively robust thermocouples. The introduction of sliding windows is shown to be effective, while the inclusion of polynomial fitting within the window produces marginal improvements in performance.
Return to index

Tu04.02: The implementation of simulated annealing combining gradient search in system identification
Click here for full paper
Authors: Mr Yiqun Zou     Dr William Heath    
Abstract: A two-stage algorithm is proposed for system identification using a maximum likelihood criterion. The first stage is a modified simulated annealing algorithm that ensures the solution avoids local minima; the algorithm is tailored for the parameter identification problem. The second stage is a standard gradient descent algorithm that ensures fast and accurate convergence to the optimum. Simulation results are presented for both linear and nonlinear system identification. The performance is compared with a breeder genetic algorithm in both cases.
Return to index

Tu04.03: Design and Real Time Implementation of Nonlinear Minimum Variance Filter
Click here for full paper
Authors: Mr. Shamsher Ali Naz     Prof. Mike Grimble    
Abstract: In this paper, the design and real time implementation of a Nonlinear Minimum Variance (NMV) estimator is presented using a laboratory based ball and beam system. The real time implementation employs a LabVIEW based tool. The novelty of this work lies in the design steps and the practical implementation of the NMV estimation technique which up till now only investigated using simulation studies. The paper also discusses the advantages and limitations of the NMV estimator based on the real time application results. These are compared with results obtained using an extended Kalman filter.
Return to index

Tu04.05: Constrained particle filtering using Gaussian sum approximations
Click here for full paper
Authors: Dr. Marc-Andre Beyer     Dr. Gunter Reinig    
Abstract: In many filtering problems, there are hard constraints in the state vector that can be a valuable source of information in the estimation process. In this contribution a method to incorporate hard state constraints in particle filters is proposed. The derived approach is based on Gaussian mixture model representation of probability distributions within the particle filter framework and a projection approach to generate constrained samples from these truncated distributions. The developed particle filters show significant improved state estimation performance and robustness against filter divergence compared to their unconstrained counterparts.
Return to index

Tu04.06: Time-delay in high-gain observer based disturbance estimation
Click here for full paper
Authors: Mr. Xuewu Dai     Prof. Zhiwei Gao     Mr. Usama Abou-Zayed     Dr. Tim Breikin    
Abstract: In this paper, the properties of a high gain observer-based disturbance estimation are analysed, and a time delay calculation approach is proposed for improving the identification of model parameter variation. The focus of this paper is the time delay between the actual disturbance and its estimate in a high gain disturbance estimation observer. It is proved, in this paper, the delay depends on the observer gain, but is independent from the model uncertainties. Thus, a novel algorithm is proposed to calculate the delay according to the phase response of disturbance estimation transfer function. The correctness of this algorithm has been verified by the simulation based on a servo motor model.
Return to index

Tu05. Fault detection and plant monitoring

Tu05.01: Discrete-time Robust Fault Detection Observer Design: a simulated annealing approach
Click here for full paper
Authors: Mr. Xuewu Dai     Mr. Yiqun Zou     Dr. Tim Breikin     Dr. Will Heath    
Abstract: The robustness in model-based fault detection has received a lot of attention during the last two decades, and RFDO (Robust Fault Detection Observer) forms an important branch of condition monitoring. However, most of current research focuses on continuous-time domain and requires relatively more computation on performance evaluation. In this paper, with the aid of the well-established eigenstructure assignment, a frequency weighted robustness index is proposed for reducing the computation costs and a left-eigenvector assignment method is presented for discrete RFDO design. A simulated annealing algorithm is applied to optimise such an observer. As illustrated in the simulation results, a better disturbance attenuation and fault detection performance have been obtained. Compared to the previous studies, simulated annealing gives the similar results as genetic algorithm, but requires a bit of less computation.
Return to index

Tu05.02: SENSOR LOCATION BASED OPTIMUM DESIGN FOR FAULT DETECTION SYSTEM
Click here for full paper
Authors: Prof. PENG Tao     Prof. Steven Ding     Prof. GUI Wei-Hua     Prof. CHEN Jie    
Abstract: This paper addresses the optimum design of fault detection systems based on sensor location. A multi-objective optimization problem based on optimal sensor location for fault detection is formulated for linear time invariant system. Measurement outputs are formed by selecting m variables of N available process measurements that ensures a high fault detetion performance. A minimum total measurement cost can be achieved when the system is designed to be as sensitive as possible to faults and simultaneously as robust as possible to the unknown inputs such as disturbance. The simulation results illustrate the effectiveness of the proposed approach.
Return to index

Tu05.03: Model-Based Sensor Fault Diagnosis in General Stochastic Systems Using LMI Techniques
Click here for full paper
Authors: Dr. Puya Afshar     Prof. Hong Wang    
Abstract: In this paper a method for sensor Fault Detection and Isolation (FDI) in non- Gaussian stochastic distribution control systems is proposed. As the output PDF is assumed measurable in probability density control methods, availability of a reliable output PDF measurements is vital. However, sensor faults occurred in practical cases can considerably affect the efficiency of the proposed PDF control algorithms. As such, studying sensor FDI in non- Gaussian stochastic distribution control systems is important. The purpose of this paper is to detect and diagnose PSD measurement sensors in a non-Gaussian system working normally under a PID control law. The proposed method is comprised of two stages, a) Nonlinear observer supervisory system design to continuously monitor Fault Detection Criteria (FDC). b) Nonlinear fault diagnosis filter design to estimate the value of the fault signal detected. Throughout the paper, the square-root PDF model has been applied and design methods are based on continuous-time Linear Matrix Inequalities (LMI) approach. Simulation results also confirm the effectiveness of the method proposed.
Return to index

Tu05.04: Multi-Agent Control of High Redundancy Actuation
Click here for full paper
Authors: Miss Jessica Davies     Dr. Thomas Steffen     Dr. Roger Dixon     Prof. Roger Goodall    
Abstract: The High Redundancy Actuator (HRA) project investigates the use of a relatively high number of small actuation elements, assembled in series and parallel in order to form a single actuator which has intrinsic fault tolerance. Both passive and active methods of control are planned for use with the HRA. This paper presents progress towards a multiple model control scheme for the HRA applied through the framework of multi-agent control.
Return to index

Tu05.05: Detection of Additive Sensor Faults in an Unmanned Air Vehicle (UAV) Model using Neural Networks
Click here for full paper
Authors: Mr Ihab Samy     Prof. Ian Postlethwaite     Prof. Dawei Gu    
Abstract: Sensor measurements are used in almost all control feedback loops and any inaccuracies can potentially lead to closed-loop instability. In this paper we make use of the online learning capabilities of neural networks (NN) to design and test a sensor fault detection and accommodation (SFDA) scheme on a nonlinear unmanned air vehicle (UAV) model. A Radial-Basis Function (RBF) neural network (NN) trained online with Extended Minimum Resource Allocating Network (EMRAN) algorithms is chosen for modelling purposes due to its good estimation capabilities and compact size. Furthermore, in an attempt to reduce false alarms (FA) and missed faults (MF) in current SFDA systems, we introduce a novel residual generator. After 47 minutes (CPU running time) of NN offline training, the SFDA scheme is able to detect additive sensor faults with zero FA and MF. It also shows good global approximation capabilities, essential for fault accommodation, with an average pitch gyro estimation error of 0.0075 rad/s.
Return to index

Tu05.06: statistical process monitoring of bioreactors: a comparison
Click here for full paper
Authors: Dr. Ognjen Marjanovic     Mr. Wu Long     Prof. Barry Lennox    
Abstract: Batch processes, such as fermenters, generally require high levels of consistency in their operation to ensure minimal losses of raw materials and product. Recent application studies have indicated that multivariate statistical technology can provide some support when trying to maintain consistent operation in complex batch processes. This paper aims to compare four different approaches to batch process monitoring using statistical methods. The comparison is made in terms of their respective ability to tolerate normal process variation while detecting abnormal operation of a process. The comparison is performed using data sets obtained from one simulated bioreactor and two industrial fermentation processes.
Return to index

Tu06. Control Theory: Optimization and design

Tu06.01: A Design-Orientated Approach to the Geometry of Fundamental Design Limitations
Click here for full paper
Authors: MR Jiqiang WANG     PROF Steve DALEY    
Abstract: There exist fundamental design limitations which explain why some designs are unattainable for the case that the performance variable is measured for feedback. For the case that performance variable is not measured it is found that new tradeoff can arise and this tradeoff is shown in this paper to have a nice geometry. The geometry, in addition to demonstrating the tradeoff, also renders itself as a design methodology. This ultimately results in a new perspective towards control design. This is remarkable since the conventional theory of fundamental design limitations explains the failure of some control designs but does not provide a specific design methodology. The design procedures are summarized for harmonic control as well as broadband control.
Return to index

Tu06.02: 'Flat Phase' PID Controllers
Click here for full paper
Authors: Dr Richard Mitchell    
Abstract: Flat Phase PID Controllers have the property that the phase of the transfer function round the associated feedback loop is constant or flat around the design frequency, with the aim that the phase margin and overshoot to a step response is unaffected when the gain of the device under control changes. Such designs have been achieved using Bode Integrals and by ensuring the phase is the same at two frequencies. This paper extends the ‘two frequency’ controller and describes a novel three frequency controller. The different design strategies arc compared.
Return to index

Tu06.03: A reduced structure controller for a Grinder Circuit system
Click here for full paper
Authors: Dr Amin Nobakhti     Prof Hong Wang    
Abstract: In this paper a reduced structure multivariable controller is developed for the model of a closed-system grinding circuit. The said controller is developed using a novel technique which employs basis pursuit regularization in order to generate a family of solutions which together span the entire range from decentralized to centralized controllers. Using this information, and the performance-cost trade off which is also computed, the designer is then able to choose the required amount of controller complexity which can achieve the desired closed-loop performance levels.
Return to index

Tu06.04: Comparison between MSE and MEE Based Component Extraction Approaches to Process Monitoring and Fault Diagnosis
Click here for full paper
Authors: Hong Wang    
Abstract: Component extraction is a technique for extracting the latent components that underlie the observation of a set of variables. In the paper both classical Principal component analysis (PCA) and autoassociative principal component neural network (PCNN) methods with minimum mean square error (MSE) criterion are compared with the corresponding extended component extraction methods with Minimum error entropy (MEE) criterion in theory. A Parzen window estimator based approximative computation method for entropy is provided, and the equivalence between MSE and MEE criteria is also analyzed. Finally, a quadruple-tank multivariate simulation example is included to evaluate the performance of the methods in process monitoring and fault diagnosis, and the encouraging results have been obtained.
Return to index

Tu06.05: Using a Chebyshev approach for the minimum-time open-loop control of constrained MIMO systems
Click here for full paper
Authors: Prof. Antonio Visioli     Mr. Stefano Piccagli    
Abstract: In this paper we propose the use of a technique based on Chebyshev polynomials approximation for determining the minimum-time rest-to-rest open-loop control law for multi-input multi-output (MIMO) continuous-time systems with input and output constraints. The optimal input can be determined, without discretising the system, by suitably approximating the state variables and the input signals by means of Chebyshev series and by subsequently solving a constrained optimizsation problem. Simulation results confirm the effectiveness of the technique.
Return to index

Tu06.06: PROPERTIES OF OUTPUT FREQUENCIES OF VOLTERRA SYSTEMS
Click here for full paper
Authors: Mr. Xingjian Jing     Dr. Ziqiang Lang    
Abstract: For a class of nonlinear systems, referred to as Volterra systems, some important properties for system output frequencies are studied in this paper. These properties demonstrate several novel frequency characteristics of system output spectrum and reveal clearly the nonlinear effects on system output spectrum from different kind and degree of nonlinearities. These new results have significance in the analysis and design of nonlinear systems or filters in order to achieve a specific output spectrum in a desired frequency band by taking advantage of nonlinearities, and provide an important guidance to applications of Volterra system theory in practices.
Return to index

Tu07. Mechatronics

Tu07.01: Fault Detection Using High Gain Observer: Application in Pipeline System
Click here for full paper
Authors: Mr Wan Rahiman     Mr Buzhou Wu     Dr Zhengtao Ding    
Abstract: This paper investigates the detection of faults in a hydraulic pipeline is presented with application of a high-gain observer for a general class of nonlinear system. Technically, the purpose of using high-gain observer is to increase the performance of the system so that it is sensitive to disturbances or faults. The adaptation scheme is literaturely straightforward and simple in the sense that the error dynamics are independent of state, input, output and unknown disturbances. Together with application results presented the effectiveness of the high-gain nonlinear observer for fault detection.
Return to index

Tu07.02: Control of a simple DC motor robot equipped with ultrasonic sensors via a field programmable gate array and a speech recognition board and microphone
Click here for full paper
Authors: Mr Andrew Tickle     Mr Paul Harvey     Dr James Buckle     Prof Jeremy Smith    
Abstract: In this paper, there is a feasibility study showing the initial development into building a practical robotic speech controlled system; this can be used to assist people with disabilities, gaining more of their independence back. Speech recognition is a very useful, and novel approach for controlling devices; the system presented here uses the 48 pin CMOS voice recognition LSI circuit, HM2007 package to perform the actual recognition for either speaker dependent or independent systems. How the Field Programmable Gate Array (FPGA) system is interfaced with the HM2007. Coding for the latter was done via the use of VHDL, and extensive Altera vector waveform analysis of the systems is shown to verify that the control and safety systems function as they were designed to. Also shown is how the FPGA is used to control the H-bridge driver chip, rather than design a system from scratch, this is due to the fact that the number of lines required would make building a logic circuit very time consuming and very complex. There is also a detailed view of how noise affects the control mechanism, how the safety features are built into the system to avoid errors and accidents from occurring, and from the system being misused by practical jokers or people with other motives.
Return to index

Tu07.03: Heuristiscs-based High-level Strategy in Multi-Agent Environment
Click here for full paper
Authors: Peter Gasztonyi     Istvan Harmati    
Abstract: In this paper, a high-level strategy concept is presented for robot soccer, based on low level heuristic inference methods, rather than explicit rule-based strategy. During tactical positioning, no strict role set is assigned for the agents, instead a fitting point of the role-space is selected dynamically. The algorithm for this approach applies fuzzy logic. We compute fields-of-quality, regarding some relevant aspects of the scenario, and integrate them into one field for each player, according to given strategic parameters (as weights). These fields will be the base of the players' decision of positioning. A significance order is also set up for the players, and their relevant location is derived from the decision-field, through subtractive clustering, in order of their significance. If an agent is in a position to manipulate the ball, an appropriate action is being selected for it. The simulation and experiments prove that the proposed approach can be efficient in dynamically changing environment or against opponents of different strategies.
Return to index

Tu07.04: STRUCTURAL ANALYSIS OF THE DEFORMATION OF THE FLEXIBLE ARM OF A ROBOT: PART 2
Click here for full paper
Authors: yaici malika    
Abstract: We consider linear structured systems in state space form where a linear system is structured when each entry of its matrices A, B, C and D are either a fixed zero or a free parameter. The structure of the system is determined by the location of the fixed zeros in these matrices. The properties of structured systems are true for almost any value of these free parameters. Theses structural properties can in general be checked by means of directed graphs that can be associated to a structured system. Its vertices correspond to the input, state and output variables, and the edges between two vertices correspond to nonzero parameters relating the corresponding variables in the equation. This paper presents an illustrative application of the previous notions. It consists in analyzing some structural properties of the deformation with linear dissipation of a flexible arm of a robot under a given acceleration modelled in state space form.
Return to index

Tu08. Adaptive Control

Tu08.01: LINEARISATION OF POWER AMPLIFIERS, USING MINIMAL CONTROL SYNTHESIS
Click here for full paper
Authors: MR BOHONG XIAO     PROF DAVID STOTEN     DR ANDREW HARRISON    
Abstract: Abstract: This report is centred on the development of a novel control algorithm for the adaptive linearisation of mobile radio frequency (RF) communications amplifiers, in order to significantly improve their distortion characteristics. The proposed adaptive linearisation methodology achieves optimal efficiency with minimal distortion. Research objectives also include the synthesis of RF amplifier dynamic models, which are generated by Saleh`s model, and the design and test of physical test systems incorporating the new forms of control.
Return to index

Tu08.02: A METHOD FOR FINDING GOOD VALUES OF ADAPTATION GAINS
Click here for full paper
Authors: Dr. Andro Rurua     Prof. Eamonn McQuade    
Abstract: Choosing the adaptation gains for a model reference adaptive controller is a complex matter. Good values of it depend on many factors in this type of non-linear system; the input amplitude and frequency, the process performance specification and the form of the controller, being the most important. This paper describes the search for good values of the adaptation gains using the MIT Rule based adaptive controller as an example. The simulations are performed using LABVIEW. The method developed is demonstrated in the real- time, speed control of a DC motor.
Return to index

Tu08.03: U-MODEL BASED ADAPTIVE INTERNAL MODEL CONTROL OF UNKNOWN MIMO NONLINEAR SYSTEMS: A CASE STUDY ON 2-LINK ROBOTIC ARM
Click here for full paper
Authors: Dr. Syed Saad Azhar Ali     Dr. Muhammad Shafiq     Dr. Jamil Bakhashwain     Dr. Fouad AL-Sunni    
Abstract: In this paper, we propose a more generalized controller design methodology for a class of nonlinear plants. This design procedure is based on MIMO U-model structure. The U-model significantly simplifies the online synthesis of the control law. The proposed technique is applied for the internal model control of a 2-link robot manipulator. The performance of the proposed U-model based internal model controller is compared to standard PID controller under different conditions.
Return to index

Tu08.04: A Disturbance Rejection Supervisor in Multiple-Model Based Control
Click here for full paper
Authors: Mr. Ehsan Peymani Foroushani     Dr. Alireza Fatehi     Prof. Ali Khaki Sedigh    
Abstract: In this paper, a multiple models, switching, and tuning control algorithm based on pole-placement control is studied. Drawbacks of the algorithm in disturbance rejection are discussed, and a novel supervisor to enhance the decision-making procedure is developed. The modified algorithm is evaluated in a simulation study for a nonlinear pH neutralization process. Comparison results are provided to evaluate the performance and robustness characteristics of the proposed algorithm.
Return to index

Tu08.05: Automatic Learning in Multiple Model Adaptive Control
Click here for full paper
Authors: Mr. Eng Ehsan Peymani Foroushani     Dr. Alireza Fatehi     Prof. Ali Khaki Sedigh    
Abstract: Control based on multiple models (MM) is an effective strategy to cope with structural and parametric uncertainty of systems with highly nonlinear dynamics. It relies on a set of local models describing different operating modes of the system. Therefore, the performance is strongly depends on the distribution of the models in the defined operating space. In this paper, the problem of on-line construction of local model set is considered. The necessary specifications of an autonomous learning method are stated, and a high-level supervisor is designed to add an appropriate model to the available model set. The proposed algorithm is evaluated in a simulated pH neutralization process which is a highly nonlinear plant and composed of both abrupt and largely continuous changes. The preference of the multiple-model approach with learning ability on a conventional adaptive controller is studied.
Return to index

Tu08.06: EXPERIMENTAL IMPLEMENTATION AND VALIDATION OF DUAL ADAPTIVE CONTROL FOR MOBILE ROBOTS
Click here for full paper
Authors: Ing. Marvin K. Bugeja     Dr. Ing Simon G. Fabri    
Abstract: This paper presents experimental results which validate the use of a novel dual adaptive controller for mobile robots operating in the presence of dynamic uncertainty. The control scheme, recently proposed by the same authors, has so far been tested by simulations only. The presented results show, for the first time, the successful application of neural network dual adaptive control in a practical mobile robot scenario. In contrast to other adaptive controllers hitherto proposed for mobile robots, the dual adaptive approach employed in this scheme does not treat estimation and control as two separate tasks, but aims to strike a balance between the two at all times. This improves the overall performance. The implementation details of the robot designed for the purpose of this research are also presented in this paper.
Return to index

Tu09. ACE: Invited session

Tu09.01: CONTROL OF INTEGRAL PROCESSES WITH DEAD TIME: PRACTICAL ISSUES AND EXPERIMENTAL RESULTS
Click here for full paper
Authors: Dr Antonio Visioli     Dr Qing-Chang Zhong    
Abstract: The problem of controlling an integral process with dead time is addressed in this paper. In particular, various practical issues concerned with the controller implementation are discussed and then verified with experiments carried out on a laboratory-scale setup where a level control problem is concerned. A comparison with a standard Proportional-Integral (PI) controller is also performed.
Return to index

Tu09.02: An Assessment of a Modified Optimal Control Strategy as Applied to the Control of an Unmanned Surface Vehicle
Click here for full paper
Authors: Dr Wasif Naeem     Prof. Robert Sutton    
Abstract: Unmanned surface vehicles (USVs) are now being used in a variety of missions including, surveillance, weapon delivery, shallow water surveying, coordination with underwater vehicles to name but a few. The performance of these unmanned systems is crucial in obtaining the required information from a given mission. The onboard navigation, guidance and control (NGC) systems, working in tandem, dictates this performance measure. Degradation in effectiveness of one system can severely affect the efficiency of the overall system. Hence the requirement of the NGC system is that of a robust type which includes fault tolerance as an integral part of the system. This paper presents results of the application of a modified optimal control strategy to an USV named Springer which has been designed and developed at the University of Plymouth for the purpose of environmental data monitoring. The performance of the proposed autopilot is compared with the standard control system in terms of real time results. Copyright © 2008 IFAC
Return to index

Tu09.03: A Fast Training Algorithm For Least-Squares Support Vector Machines
Click here for full paper
Authors: Miss Xiao Lei Xia     Dr Kang Li     Prof Minrui Fei    
Abstract: The paper addresses the issue of training acceleration for binary Least Squares Support Vector Machines (LS-SVMs). An LS-SVM is trained by solving a linear system, for which the conjugate-gradient (CG) method is applied however in a very complicated way, therefore slows down the training process. To overcome the drawback, this paper first introduces a variant formulation of LS-SVMs to achieve explicit application of the CG method. Then, an alternative to the CG method - namely Forward Linear Regression (FLR) is proposed to further speed up the training process. Different from the CG method, the FLR, derived from a two-stage algorithm for fast model selection in nonlinear system identification, is a non-iterative algorithm and is easy to implement. Experimental results on the two-spiral dataset confirm the efficacy of the proposed techniques.
Return to index

Tu09.04: INTELLIGENT SOFTWARE SENSORS FOR FED-BATCH FERMENTATION PROCESSES
Click here for full paper
Authors: Dr Hongwei Zhang    
Abstract: Software sensors have attracted great research interests due to the problem of lacking suitable and robust online sensors for key fermentation variables in fed-batch fermentation processes. In this paper, intelligent software sensors have been developed based on multivariate statistical process control methods. The software sensors not only provide real time estimation of key variables but also have the facilities of self-diagnosis, self-validation and self-calibration using available lab assay data. An application of the software sensors to a fed-batch penicillin fermentation process is presented, and significant improvements over ordinary methods have been shown in the simulation results.
Return to index

Tu09.05: An economic parametrisation for parahermitian matrix functions used in control systems optimisation
Click here for full paper
Authors: Dr Alexander Lanzon    
Abstract: Positive parahermitian matrix function descriptions occur frequently in optimisation problems that arise in control theory. Parahermitian matrix functions can however be parametrised in a number of different equivalent ways. This brief note discusses an economic parametrisation which leads to substantially less variables that are needed in optimisation.
Return to index

Tu09.06: Robust Control of a High Redundancy Actuator
Click here for full paper
Authors: Dr Thomas Steffen     Dr Roger Dixon     Prof Roger Goodall     Dr Argyrios Zolotas    
Abstract: The High Redundancy Actuator project deals with the construction of an actuator using many redundant actuation elements. Whilst this promises a high degree of fault tolerance, the high number of components poses a unique challenge from a control perspective. This paper shows how a simple robust control can be used to control the system both in nominal state and after faults. To simplify the design task, the parameters of the system are tuned so that a number of internal states are decoupled from the input signal. If the decoupling is not exact, there may be small deviation from the nominal transfer function, especially when a fault has occurred. The robustness analysis ensures that the system performs well for all expected behaviour variations.
Return to index

Tu10. Predictive Control

Tu10.01: A Model Predictive Approach to Wireless Networked Control
Click here for full paper
Authors: Mr Jian Chen     Prof. George Irwin     Mr Adrian McKernan     Dr William Scanlon    
Abstract: In Wireless Networked Control Systems (WNCS), time-varying and unknown delays can significantly degrade the closed-loop performance and even lead to instability. This paper proposes using Dynamic Matrix Control for WNCS, where, a set of predicted control signals corresponding to possible delays are sent to the plant. Here the most appropriate control signal is selected based on the round-trip delay as a QoS measure of the wireless channel condition. Results from Monte Carlo simulations on a cart-mounted inverted pendulum confirm the efficacy of the method. Here the random delay introduced by WNCS are modelled by an Inverse Gaussian distribution derived from experiments on an IEEE 802.11b network.
Return to index

Tu10.02: Design of Reconfigurable Predictive Control Applied on the Air Path of a Diesel Engine
Click here for full paper
Authors: Ms Layerle Khaoula     Dr Langlois Nicolas     Dr Chafouk Houcine    
Abstract: In this paper, a method for reconfigurable predictive control of the air path of a Diesel engine system is presented. Failures are identified indirectly by estimating parameters of the linear engine model using the recursive least square algorithm (RLS). The actuators of the air system considered here are variable geometry turbine (VGT) and exhaust gas recirculation valve (EGR). The aim of the reconfiguration controller is to track simultaneously the desired trajectories of intake pressure (P1) and exhaust pressure (P2) when a faults occurs. Some simulation results are presented and compared to GPC applied on coupled MIMO system. The proposed controller exhibits good control performance: it ensures global stability and tracking of output references ithout zero offset. Moreover, the separated optimization of the GPC parameters for each subsystem permits the controller to have good performance during transient mode especially in terms of overshoots.
Return to index

Tu10.03: Model Predictive Control of Substructured Systems
Click here for full paper
Authors: Dr Guang Li     Miss Jia-Ying Tu     Prof. David Stoten    
Abstract: In this paper, we consider the control of multivariable substructured systems with input constraints. Model Predictive Control (MPC) is used to synchronize the interface between the physical and numerical substructures. As a case study, a quasi-motorcycle suspension system is converted into a multivariable substructured system. An MPC controller is developed for this system. Simulation results show the advantage of using an MPC controller to synchronize the substructured system.
Return to index

Tu10.04: Constrained Predictive Control Of A Servo-driven Tracking Turret
Click here for full paper
Authors: Dr. Peter Martin     Dr. Nick Brignall     Mr Matt MacDonald     Prof. Mike Grimble    
Abstract: Vehicle-mounted 2-axis turrets are widely used in high bandwidth tracking systems, frequently encountered in air-to-ground, ground-to-air and air-to-air targeting. Existing controllers for these systems are generally implemented in classical PID form. The objective of this paper is to examine the novel application of constrained model predictive control (MPC) to a Selex turret simulation. The characteristics of the control problem are well matched to MPC, as hard saturation constraints are present in the electrical subsystem and a reference trajectory can be generated for several seconds in advance due to the predictability of a missile trajectory. The state-space model and Kalman filter are described, and simulation results are presented to demonstrate the validity and superior performance of the MPC method.
Return to index

Tu10.05: AQM Control of TCP/IP Networks using Generalized Predictive Control
Click here for full paper
Authors: Dr Teresa Alvarez     Dr. Smaranda Cristea    
Abstract: This paper presents how generalized predictive control can improve the performance of TCP/IP networks when dealing with control congestion. It is shown that predictive control and GPC (Generalized Predictive Control), in particular, can be seen as an improved AQM (Active Queue Management) method. Predictive controllers, constrained and unconstrained, are compared with other control methods, such as PI control or RED/AQM, showing the advantages of the proposed technique, as it makes teh consideration of constraints possible in the manipulated and controlled variables.
Return to index

Tu10.06: Subspace-based Model Predictive Control with Data Prefiltering
Click here for full paper
Authors: Mr. Noor Azizi Mardi     Prof. Liuping Wang    
Abstract: Subspace-based model-free predictive control algorithms directly estimate the relevant components of a predictive controller. Due to disturbances and noise in the measured data, the estimation results were often poor, which limited the applications of subspace-based model-free predictive controllers. By assuming a priori knowledge of the disturbance characteristics, this paper proposes a subspace-based model-free predictive control algorithm that utilizes the noise model for the estimation of the predictive control gain matrices. Simulation results show improved control results.
Return to index

Tu11. Control Theory: Optimization and nonlinear design

Tu11.01: Controller Design of Conflict Multi-objective control problem by Preference
Click here for full paper
Authors: Phd zhu bingkun     Pro. xu lihong    
Abstract: According to the character that the optimal point is not single in the conflict Multi-Objective Control Problem (MOCP) and optimal solutions cannot be simultaneously obtained by traditional optimization methods in a single simulation run, a new algorithm based on evolutionary computation is presented, which incorporates user’s preference information into optimal process for obtaining dense Pareto solutions in preference region and defines a new selection function making control objectives stabilized in this region.
Return to index

Tu11.02: Control Engineering at the University of Manchester in the Post War Years
Click here for full paper
Authors: Prof Derek Atherton    
Abstract: The purpose of this paper is to outline the leading role the Electrical Engineering Department at Manchester played in the early development of control engineering teaching and research in the UK after the war. The holding of a major UK conference in Control Engineering in Manchester roughly 60 years after the start of this work seems an appropriate time to remind today's researchers of these early contributions and the changes in technology and the university environment.
Return to index

Tu11.03: An Arnoldi Based Method to Discrete Time Linear Optimal Multi-periodic Repetitive Control
Click here for full paper
Authors: Mr Youde Han     Prof David H Owens    
Abstract: For LTI plant, a benchmark tracking solution was recently proposed by optimal technique for discrete time linear multi-periodic repetitive control system that gives asymptotic perfect tracking, where the original tracking problem was transfered into a regulator problem by developing a new state-space representation that combines the plant and demand signal. However, in practice, the periods of the demand signals are usually very large, therefore the dimension of this new plant description increases and that naturally leads to high order computations for solving discrete Riccati equation. In order to overcome this problem, an Arnoldi based method is applied in this paper to first reduce the high order state-space representation to a low order counterpart, and then a direct method is used to solve its corresponding low order riccati equation. Finally a projection between two riccati solutions is applied to retrieve an approximate high order riccati solution. This result also leads to a new multi-periodic repetitive controller in terms of low order riccati solution. A numerical example is given and asymptotic perfect tracking is guaranteed.
Return to index

Tu11.04: A non-parametric method for nonlinear J-J' spectral factorisation.
Click here for full paper
Authors: Dr Andrew Shenton    
Abstract: This paper presents a Newton-iteration method for obtaining a J-J$'$ spectral factorisation of systems from non-parameteric characterisations using identification techniques.Systems are assumed to be Fr\'{e}chet differentiable discrete-time maps. The technique may be used on nonlinear nonparametric time-response representations. The scheme requires stabilised identification and stabilised inverse identification and with such can be used to iterate on NARMAX controller structures.
Return to index

Tu11.05: Design of Retarded Fractional Delay Differential Systems by the Method of Inequalities
Click here for full paper
Authors: Dr Suchin Arunsawatwong     Mr Van Quang Nguyen    
Abstract: Methods based on numerical optimization are useful and effective in the design of control systems. This paper describes the design of retarded fractional delay differential systems (RFDDSs) by the method of inequalities, in which the design problem is formulated so that it is suitable for solution by numerical methods. This is an extension of the formulation proposed by Zakian and Al-Naib (Proc. IEE 120, pp. 1421--1427, 1973) in connection with rational systems. In using the formulation with RFDDSs, the associated stability problems are resolved by using the stability test and the numerical algorithm recently developed by the authors, whereas the time-responses are obtained by using a known method for numerical inversion of Laplace transforms. Two numerical examples are given, where fractional controllers are designed, respectively, for a time-delay plant and a heat-conduction process.
Return to index

Tu12. Decision and control (Invited)

Tu12.01: Decision-Making of Football Agents with Support Vector Machine
Click here for full paper
Authors: Dr. Hisashi Handa     Mr. Satoshi Kajiyama    
Abstract: Robocup has attracted much attention for Artificial and Computational Intelligence researchers. Robocup involves various aspects of problems, i.e., coorporation with team mates, dynamic problems, imperfect information, uncertainness caused by noise, and so on. Therefore, it is quite difficult to design football agents. In this paper, Support Vector Machines, one of the most famous machine learning algorithms, are used to decide if the agents carry out basic skills, such as shoot and through balls, which are given in advance. That is, firstly, data, i.e., the position and directions of balls and players, is collected by playing given skills naively. Then, labels indicating the success/fault of the skills are added to the data. Secondly, SVM learns the data. Finally, the SVM decides if the skill should be carried out. Several experiments on game plays with stronger team binaries at Japan Open elucidate the effectiveness of the proposed method.
Return to index

Tu12.02: OPERATOR BASED ROBUST NONLINEAR CONTROL SYSTEM DESIGN OF MIMO NONLINEAR FEEDBACK CONTROL SYSTEMS
Click here for full paper
Authors: Dr. Shuhui Bi     Dr. Mingcong Deng     Dr. Akira Inoue    
Abstract: In this paper, operator based robust nonlinear control system design of a multi-input multi-output nonlinear feedback control system is proposed, that is, robust stability of the MIMO system is studied by using operator based robust right coprime factorization approach. Some sufficient conditions for the MIMO nonlinear systems to be robust stable are derived. As a result, robust nonlinear control system is designed for the MIMO system. Final, an example is given to initially demonstrate the theoretical analysis.
Return to index

Tu12.03: LS-SVM based motion control of a mobile robot in dynamic environment
Click here for full paper
Authors: Dr. L. Jiang     Dr. M. Deng     Dr. A. Inoue    
Abstract: In this paper, least squares support vector machine (LS-SVM) based motion control of a mobile robot in dynamic environment is proposed under the measured data with uncertainties. The proposed scheme can control the robot by consideration of local minima, where the controller is based on Lyapunov function candidate and considers virtual forces information. Comparing with standard support vector machine (SVM) method, LS-SVM method is used for estimating the control parameters from the measured data with uncertainties. Simulation results are presented to show the effectiveness of the proposed scheme.
Return to index

Tu12.04: OPERATOR BASED FAULT DETECTION SYSTEM DESIGN TO AN ACTUATOR FAULT OF A THERMAL PROCESS
Click here for full paper
Authors: Dr. Mingcong Deng     Dr. Akira Inoue     Dr. Kazunori Edahiro    
Abstract: This paper proposes a fault detection method for an actuator fault of an aluminum plate thermal process with input constraints. Operator-based robust right coprime factorization approach is utilized in this method. In details, after creating a mathematical model, a robust tracking operator system is designed. Following this, design of the fault detection system is given. Finally, experiment is conducted to support the proposed design method.
Return to index

Tu12.05: COMARISION OF CONTROL DESIGN TECHNIQUES FOR A NUCLEAR REACTOR
Click here for full paper
Authors: Mr Zakwan Skaf     Prof Hong Wang    
Abstract: This paper presents a comparative study of different control design methods applied to a nuclear reactor. A nuclear reactor temperature controller is designed using the H-infinity control. This advanced controller is compared with a conventional PID controller and a traditional optimal controller design using LQG method.
Return to index

Tu13. Fuzzy Control Systems

Tu13.01: A new engineering method for fuzzy reliability analysis of surge detection and isolation in centrifugal compressor
Click here for full paper
Authors: HAFAIFA Ahmed     LAAOUAD Ferhat     LARAOUSSI Kouider    
Abstract: A new dual fuzzy controller for the nonlinear model of the compression system is proposed in this paper. The surge phenomenon in the centrifugal compressor, the non-linearities and uncertainties of the compression system make it impossible to use a conventional controller over a wide range of operation. This fuzzy controller is designed to consist of a active surge control and phase control without any explicit system models, but driven in the human thinking mechanism. A simulation example of compression system is given to demonstrate the validity of the proposed control scheme. It is shown that the fuzzy controller can be simplified, and good tracking control performance can be achieved by choosing appropriate fuzzy roles. However, the dual fuzzy controller can successfully intervene in the control surge of the compression system. This new fuzzy control methodology suggested in this work reproduced well the main characteristics of the turbo compressor dynamic model developed by Moore and Gretzer and give place to a more precise and easy to handle representation. It is about a inaccuracies reproducing with a certain degree of satisfaction of the real process without being as much complex.
Return to index

Tu13.02: MODELLING AND CONTROL OF FES-ASSISTED INDOOR ROWING EXERCISE
Click here for full paper
Authors: Mr Zakaria Hussain     Dr M Osman Tokhi     Dr C Samad Gharooni     Mrs Siti Fauziah Toha    
Abstract: This paper present the development of a model of indoor rowing exercise for paraplegics. Indoor rowing exercise is introduced as a hybrid exercise for restoration of function of lower extremities for paraplegics through the application of functional electrical stimulation (FES). Two stimulated muscle model, quadriceps and hamstrings are developed for knee extension and flexion. A novel fuzzy logic control strategy is designed to control the rowing manoeuvre. Simulation results verifying the control strategy are presented and discussed.
Return to index

Tu13.03: Adaptive Fuzzy Model-based Predictive Control Using Fuzzy Decision Making
Click here for full paper
Authors: Dr. Yue Wu     Prof. Arthur Dexter    
Abstract: Motivated by the need to develop more effective methods of controlling uncertain non-linear systems, this paper focus on developing an adaptive fuzzy model-based controller, in which the optimisation variables remain in fuzzy domain. The scheme uses an on-line fuzzy identification scheme, which is able to generate a fuzzy relational model using the training data from the system. The proposed control system is applied to the supply air temperature control in a simulated cooling coil system of an air-conditioning system to evaluate the improvement of the proposed scheme compare to the non-adaptive version of the controller.
Return to index

Tu13.04: FUZZY CONTROLLER DEVELOPMENT FOR A PEM FUEL CELL SYSTEM
Click here for full paper
Authors: Dr Kary Thanapalan     Prof Guoping Liu     Mr Jonathan Williams     Dr David Rees    
Abstract: A Polymer Electrolyte Membrane (PEM) fuel cell system model that is suitable for control study is presented in this paper. The PEM mathematical model is then used for the controller development to improve system performance. Within the University research facilities, there is a PEM Fuel Cell Test station (PEM –FCT) available, so the PEM-FCT is used for the modelling and controller study. A fuzzy set-point weighted PID controller is designed to improve the performance of the fuel cell system. The underlying idea of our controller design is to use a fuzzy based system to support the operation of a PID controller. The new control strategy is implemented on a PC based computer model of the FCT system and simulated. The results indicate that the control strategy has improved the system performance dramatically.
Return to index

Tu13.05: Reinforcement Learning for Probabilistic Fuzzy Controllers
Click here for full paper
Authors: Mr William Hinojosa     Dr Samia Nefti     Prof John Gray     Prof Uzay Kaymak    
Abstract: In this paper a new hybrid approach combining Reinforcement Learning and a Probabilistic Fuzzy controller is proposed. This structure is based on a reinforcement learning agent that measures the performance of a system and uses this to reinforce and adapt the rule base and related probabilities in order to achieve its goal. The proposed reinforcement learning algorithm is based on a modified version of the actor-critic architecture for dynamic reactive compensation. Experiments based on simulations using a DC motor numerical model were carried out in order to validate the proposed approach. The obtained numerical results show that our proposed algorithm outperforms the classic Reinforcement Learning in term of learning time and accuracy.
Return to index

Tu14. Systems control using neural networks

Tu14.01: Controller Design for Nonlinear Systems with Stochastic Time Delays Using Neural Networks and Information Entropy
Click here for full paper
Authors: Dr. J.H. Zhang     Dr. A.P. Wang     Dr. Hong Wang    
Abstract: In this paper, a novel controller is proposed for unknown discrete-time nonlinear systems with uncertain output-channel time delays using RBF neural networks and information entropy. The controller is designed by minimizing the quadratic Renyi entropy. The probability density function (PDF) of the closed loop tracking error is estimated by Parzen windowing technique, where the Jacobian information of the system is estimated by an RBF neural network. The convergent condition of the proposed control algorithm is given. A simulation example is included to show the effectiveness of the proposed algorithm.
Return to index

Tu14.02: Controller Design of Nonlinear TITO Systems with Uncertain Delays via Neural Networks and Error Entropy Minimization
Click here for full paper
Authors: Dr. J.H. Zhang     Dr. A.P. Wang     Dr. Hong Wang    
Abstract: In this paper, a novel control algorithm for nonlinear two input and two output (TITO) systems with random input and output delays is presented. Due to the stochastic characteristics induced by uncertain time delays, TITO feedback control systems are cast into a general framework, where the controllers are designed based upon minimizing the entropies of tracking errors. The controllers that have been implemented by BP neural networks are obtained without decoupling. The convergence in the mean square sense is analyzed. Simulation results show the effectiveness of the proposed approaches.
Return to index

Tu14.03: Single Network Adaptive Critic for Vibration Isolation Control
Click here for full paper
Authors: Dr. Jia Ma     Dr. Tao Yang     Dr. Zeng-Guang Hou     Dr. Min Tan    
Abstract: Vibration isolation control is the critical issue to guarantee the performance of various vibration-sensitive instruments and sensors in practical engineering systems. In this paper, single network adaptive critic (SNAC) based controllers are developed for vibration isolation applications. The SNAC approach differs from the typical action-critic dual network structure in adaptive critic designs (ACDs) by eliminating the action network, which leads to substantial computational savings. Two training methods, i.e., the off-line and online methods are proposed to adapt the SNAC controllers respectively. In contrast with the existing off-line SNAC training method, the off-line method proposed in this paper adopts the least mean square (LMS) training algorithm with variable learning rate to make the training procedure converge faster. Furthermore, for real-time control purpose, the online learning method is presented for tuning the weights of the critic networks along the real-time state trajectories of the isolation system. Additionally, the ``shadow critic" training strategy used in the online method further improves the convergence rate. Simulation results have shown that the developed SNAC controllers using the different training methods can converge to the continuous-time optimal control solution at satisfactory speed. Moreover, the designed SNAC controllers alleviate vibration disturbance more effectively and have better control performance in comparison with the passive isolator.
Return to index

Tu14.04: Discrete-Time Decentralized Neural Identification and Control for a 2 DOF Robot Manipulator
Click here for full paper
Authors: M.C. R. Garcia-Hernandez     Dr. E.N. Sanchez     Dr. A.G. Loukianov     Dr. E. Bayro-Corrochano     Dr. V. Santibañez    
Abstract: This paper presents a discrete-time decentralized control scheme for identification and trajectory tracking of a 2 DOF robot manipulator. A recurrent high order neural network (RHONN) structure is used to identify the plant model and based on this model, a discrete-time control law is derived, which combines discrete-time block control and sliding modes techniques. The neural network learning is performed on line by Kalman filtering. A controller is designed for each joint, using only local angular position and velocity measurements. These simple local joint controllers allow trajectory tracking with reduced computations. The applicability of the proposed scheme is illustrated via simulations.
Return to index

Tu14.05: Neural Predictive Control for Wide Rage of Process Systems
Click here for full paper
Authors: Mr. Seyed Ali Jazayeri Moghadas     Dr. Alireza Fatehi     Dr. Houman Sadjadian     Prof. Ali Khaki-Sedigh    
Abstract: In this paper a Neural Predictive Controller (NPC) designed to control wide range of process systems. Neural Network identification yields nonlinear global map of the unknown system. Levenberg-Marquardt (L-M) optimization method is used to find optimal control signal to minimize future errors of the objective function of predictive controller. Inequality constraints of actuators are added to the objective function through a penalty term which increases drastically as it approaches limitations. To use the controller for wide range of process systems, an initial phase runs before the main controller to determine parameters. This phase moves the system output to operating point and applies PID controller with APRBS reference signal. The gathered data are used to estimate parameters such as pure delay, prediction horizon, control signal term coefficient and identification order. To validate the approaches, the controller has implemented in level, pressure and flow plants and compared with conventional controller which shows faster and smoother tracking results.
Return to index

Tu15. Modeling and Simulation

Tu15.01: MODELLING, PARAMETER ESTIMATION AND VALIDATION OF A 300W PEM FUEL CELL SYSTEM
Click here for full paper
Authors: Dr Kary Thanapalan     Mr Bo Wang     Mr Jonathan Williams     Prof Guo-Ping Liu     Dr David Rees    
Abstract: In this paper a 300W PEM FC Stack dynamic model is developed and implemented in MATLAB/Simulink. Using semi-empirical equations for modelling a Proton Exchange Membrane (PEM) fuel cell is proposed for providing a tool for the design and analysis of fuel cell stack systems. The modelling results are compared with experimental results. The comparison shows good agreements between the modelling results and experimental data. The model could be used in PEM fuel cell control related studies.
Return to index

Tu15.02: Modelling and Parameter Identification of Electrochemical Cu-Cu Cell
Click here for full paper
Authors: Mr Alexander Mendelson     Dr Robert Tenno    
Abstract: An electrochemical cell consisting of two copper electrodes and copper sulphate solutions is modeled. The presented model takes into account both electrode interfaces and the activity of copper(II) ions. The current-potential equation is derived and a method for identifying mass-transfer parameters as well as kinetic parameters is proposed for a specific case. The created model is simulated and verified against measurements.
Return to index

Tu15.03: Phase Model for the relaxed van der Pol oscillator and its application to synchronization analysis
Click here for full paper
Authors: Dr. Joaquin Collado     M.C. Olivia Mimila-Prost    
Abstract: A one dimensional phase model for the classic two dimensional van der Pol oscillator is developed. This model is restricted to the relaxed case, and its construction is based on the slow and fast transitions the phase goes through during its cycle. An application of the phase model is included in which synchronization of two coupled van der Pol oscillators is analyzed and even used to calculate the coupling strength needed for their synchronization.
Return to index

Tu15.04: FPGA IMPLEMENTATION OF WHEEL-RAIL CONTACT LAWS
Click here for full paper
Authors: Mr. Yongji Zhou     Dr. T.X. Mei     Dr. Steven Freear    
Abstract: This paper presents the development of an accelerator for the real-time simulation of wheel-rail contact laws (Hertz and Fastsim), which would enable the use of hardware-in-the-loop for experimental studies of latest active control technology for wheelset stabilization and steering. The complex wheel-rail contact laws are implemented using a single FPGA chip that outperforms modern general-purpose CPU or DSP in the aspects of processing time, configuration flexibility and cost. Fastsim algorithm is restructured to utilize FPGA's parallel processing feature. Reusable IP cores (FPU) are used for the floating point operations. The scheduling of the operations is optimised to ensure effective and efficient allocation of the FPGA's resources.
Return to index

We01. Control Applications: Low complexity control

We01.01: Study of Reduced-order and Non-linear Local Optimal Control Application to Aero Gas Turbines
Click here for full paper
Authors: Zukhra Kamalova     Mahmoud Ashry     Tim Breikin    
Abstract: In this work second order linear model, reduced order (first order) linear model and second order non-linear model of a gas turbine engine have been obtained from the engine input-output data using evolutionary optimization technique. These three models have then been used in local optimal control design. The obtained controllers have been applied to the second order non-linear engine model and their performance has been compared.
Return to index

We01.02: Equalisation Tuning Method
Click here for full paper
Authors: Dr. Damir Vrancic     Prof. Raymond Gorez     Prof. Stanko Strmcnik    
Abstract: The paper presents a novel tuning method for different types of controllers. Since the tuning method is trying to equalise the closed-loop response to the open-loop response, it is named “Equalisation tuning method”. The main advantage of this method is that it does not require any additional data from the operator except the measurement of the process steady-state change in an open-loop experiment. The equalisation method is also relatively insensitive to process output noise. Simplicity and efficiency of the method is demonstrated on several process models and on a hydraulic laboratory plant. Matlab and Simulink files are provided.
Return to index

We01.03: Low Complexity Control of Hybrid Systems with Application to Control of Step-down DC-DC Converters
Click here for full paper
Authors: Dr. Jalal Habibi     Prof. Behzad Moshiri     Prof. Ali Khaki Sedigh    
Abstract: Control of hybrid systems as those systems with mixed time-driven and event-driven dynamics faces the computational complexity as a main challenging problem. Explicit solution to the optimal control problems has been proposed as a tool to reduce the on-line computational burden. The complexity of the explicit solution is again prohibitive for large problems. This paper shows that how a recently-proposed approach by the authors can be utilized to reduce the computational complexity in explicit predictive control of hybrid systems. The proposed approach generates a family of suboptimal controllers for which the complexity and error can be controlled by a tuning parameter. The closed loop stability is guaranteed by a contractive constraint and is preserved in all suboptimal controllers. Application of the proposed scheme to hybrid control of synchronous step-down DC-DC converters clarifies the steps for modeling and controller design as well as the achieved computational benefits.
Return to index

We01.04: A Comparative study on charge system modelling in fine paper production
Click here for full paper
Authors: Dr. Puya Afshar     Prof. Hong Wang     Mr. Neil Strain    
Abstract: This application-oriented paper provides a comparative study on modelling methods with application to a functional fine paper production machine. The aim is to develop a model for the charge measurement system in wet-end paper making systems for future control purposes. However, a Multi Input Multi Output (MIMO) model has been proposed to further generalise the proposed model. A series of six-month worth machine's input-output data are employed to develop different models. The three models, namely linear, dynamical Neural Network (NN), and a so-called hybrid model are developed to model the paper machine's behaviour. The hybrid model consists of a dynamical linear part and dynamical NN part. The linear part will model the machine around each operating point. The dynamical NN part will help to extract further un-modeled nonlinearities. Simulation results and variety of validation tests confirm that the hybrid model can effectively represent the paper machine dynamics.
Return to index

We01.05: Robust H-Infinity Control of a Steerable Marine Radar Tracker
Click here for full paper
Authors: Dr Stuart Crawshaw     Dr Daniel Auger     Mr Stephen Hall    
Abstract: This paper describes the application of a robust control technique to a steerable marine radar tracker intended to provide good performance with minimum operator intervention over the course of its lifecycle. The paper shows that the sightline steering problem can be decoupled from the target observation problem and uses a well-known robust control technique ($H_{\infty}$ loop-shaping) to synthesize a controller. Analytic bounds on the stability of the closed loop are stated by considering a model set parameterized on Vinnicombe’s nu-gap metric. Experimental verification exercises are briefly described, and proposals for formal validation work using interpolation techniques are made.
Return to index

We01.06: IDENTIFICATION OF PARKINSON’S DISEASE TREMOR ONSET USING ARTIFICIAL NEURAL NETWORKS
Click here for full paper
Authors: Prof Kevin Warwick     Dr Mark Gasson     Mr Jon Burgess     Mr Song Pan    
Abstract: In this paper we consider the possibility of using an artificial neural network to accurately identify the onset of Parkinson’s Disease tremors in human subjects. Data for the network is obtained by means of deep brain implantation in the human brain. Results presented have been obtained from a practical study (i.e. real not simulated data) but should be regarded as initial trials to be discussed further. It can be seen that a tuned artificial neural network can act as an extremely effective predictor in these circumstances.
Return to index

We02. Control Applications : Optimization and Networks

We02.01: Nonparametric Collocation ODE Parameter Estimation: Application in Biochemical Pathway Modelling
Click here for full paper
Authors: Mr. Fei He     Dr. Martin Brown     Mr. Choujun Zhan     Dr. Lam Fat Yeung    
Abstract: Parameter estimation of non-linear differential equations has long been an active and challenge research area. Conventionally methods are computationally intensive and often poorly conditioned. In the context of biochemical pathway modeling, a new method focused on this paper is the so-called "collocation" method, which is a nonparametric data smoothing based approach. The statistical property of a sort of linear smoothing spline based collocation methods is explicitly analyzed. It is concluded that this approach is computational efficient, but leads a non-zero estimation bias and it changes the independence assumption in the additive noise.
Return to index

We02.02: Load Minimization Design for Internet-based Control
Click here for full paper
Authors: Prof. Shuang-Hua Yang     Mr Z Zhang     Ms Y Li     Prof. Q-G Wang    
Abstract: This paper presents a design method for Internet-based control systems in a dual-rate configuration to achieve load minimization and dynamic performance specifications. It avoids the complexity of large scale system design by focusing on individual control systems. In the dual-rate configuration, the plant under control is first stabilized by a local controller with a high sampling rate. The remote PID controller, which regulates the output according to the desirable reference, adopts a low sampling rate to reduce load on the network. The upper bound of the remote PID controller's sampling time which meets the requirement on control performance is derived and a simple tuning method for the remote PID controller is presented. Simulation and real-time examples are provided for illustration.
Return to index

We02.03: Adaptive Feedforward Control via Virtual Error Approach with Application to Predistortion of Nonlinear HPA
Click here for full paper
Authors: Prof. Akira Sano     Mr. Tomohiro Ohno    
Abstract: A novel virtual error approach is proposed for fully adaptive feedforward control, which plays an importnat role in nonlinear active noise control and predistortion for nonlinear high power amplifier (HPA). To attenuate the compensation error, two kinds of virtual error are introduced and are forced into zero by adjusting three nonlinear adaptive filters in an on-line manner. It is shown that the convergence of the compensation error to zero can be assured by forcing the virtual errors to zero separately. The proposed method can adjust the predistorter directly without identification of a post-inverse model of HPA as adopted in previous predistortion methods. The effectiveness of the proposed virtual error approach is clarified in comparison with an ordinary nonlinear filtered-x algorithm in the adaptive predistortion for nonlinear HPA used in OFDM communication systems.
Return to index

We02.04: Design and Implementation of Brushless Motor Controller Based on SOPC
Click here for full paper
Authors: Stu. liu qingqiang     Pro. Qi Hui    
Abstract: In this paper, a kind of design of BLDM (Brushless DC motor) controller based on SOPC (System on programmable chip) is introduced. CPU, BLDM switching, PWM generator, and data acquisition module are integrated in a single FPGA chip. This design improves the integration, anti-interference and makes the system easy to promote. Experiment result has proved that the steady and dynamic performance of BLDM controller based on SOPC is so good meet the requirement of servo-system.
Return to index

We02.05: Motion stabilization in the presence of friction and backlash: a hybrid system approach
Click here for full paper
Authors: Dr Lorinc Marton     Dr. Bela Lantos    
Abstract: In this paper a hybrid system approach is considered to deal with backlash and friction induced nonlinearities in mechanical control systems. To describe the low velocity frictional behaviour a linearized friction model is proposed. The novelty of this study is that based on the introduced friction model, the stability theorems developed for hybrid systems can directly be applied for controller design of mechanical systems in the presence of Stribeck friction and backlash. During the controller design it is assumed that the size of the backlash gap is unknown and the load side position and velocity cannot be measured. For motion control an LQ controller is applied. A condition is formulated for the control law parameters to guarantee the asymptotic stability of the control system. Simulation measurements were performed to confirm the theoretical results.
Return to index

We03. Advanced Process Control

We03.01: Modeling and Control of a Fluidised Bed Dryer
Click here for full paper
Authors: Dr Javier Villegas     Dr Stephen Duncan     Dr Haigang Wang     Prof Wuqiang Yang     Mr Rambali Raghavan    
Abstract: In this paper, the modeling and control of the moisture content of the particles in a batch fluidised bed dryer are studied. First, a lumped mechanistic model is developed to describe the heat and mass transfer between solid, gas and bubble phases and experimental validation shows that the model can be used to predict the particle moisture content and temperature profiles during the drying process in the bed dryer. Feedback control of material moisture content in a bed dryer is studied where the moisture content is obtained by measuring the humidity and temperature of the outlet gas. A controller is designed to achieve a desired drying rate for wet materials.
Return to index

We03.02: MEMBRANE MODELING FOR SIMULATION AND CONTROL OF REVERSE OSMOSIS IN DESALINATION PLANTS
Click here for full paper
Authors: Dr. Fernando Tadeo     Dr. S. Syafiie     Mr. Luis Palacin     Prof. Cesar de Prada    
Abstract: A mathematical model of Reverse Osmosis membranes is proposed for use in the testing and comparison of control strategies in Reverse Osmosis plants. The model has been developed so that it can be used within off-the-shelf software and the parameters are simple to obtain from available plant measurements. Some simulations of the proposed model show that it correctly reproduces the expected process responses and can be used for testing different control strategies.
Return to index

We03.03: CONTROLLING WATER QUALITY USING REVERSE OSMOSIS: THE DEVELOPMENT OF SIMPLIFIED DYNAMIC MODEL
Click here for full paper
Authors: Dr. Mohammad Al-haj Ali     Prof. AbdulHamid Ajbar     Prof. Khalid Alhumaizi     Prof. Emad Ali    
Abstract: Reverse osmosis (RO) is a compact process for the removal of ionic and organic pollutants from contaminated water. This study deals with the development of dynamic model for tubular reverse osmosis unit. The proposed model describes the unit as a series of single tubes, each tube is described by two ordinary differential equations (ODE) and the whole module is to be described by sets of differential algebraic equations. The tubes are modeled and solved sequentially where the output of any tube becomes the input for the next one. The predictions of steady state and dynamic models are in good agreement with the experimental results of a lab scale RO unit. This model is simpler than the currently used distributed models; besides it gives more insight than black box models. This model can be used to improve the understanding of RO processes as well as to develop different model-based control algorithms for this process.
Return to index

We03.04: Noncausal open-loop control with combined system identification and PID controller tuning
Click here for full paper
Authors: Prof. Antonio Visioli     Dr. Claudio Carnevale     Prof. Aurelio Piazzi    
Abstract: In this paper we propose a systematic methodology that integrates the three main phases of the design of an industrial control system, namely, the identification phase, the tuning of the (PID) feedback controller and the design of a (noncausal) open-loop action. In particular, a first-order-plus-dead-time model of the process is estimated after having filtered properly the data collected in the identification experiment. Then, the tuning of the controller is based on frequency loop shaping where the target closed-loop system bandwidth is selected by considering the desired output transition time from one set-point value to another. Finally, the noncausal open-loop command input is synthesised by applying a stable input-output inversion procedure. Simulation results show the effectiveness of the methodology.
Return to index

We04. Control Applications: Automotive

We04.01: Fuel consumption optimization for a city bus
Click here for full paper
Authors: Dr NOUVELIERE Lydie     Dr BRACI Mohamed     MENHOUR Lghani     LUU Hong Tu     Dr MAMMAR Saïd    
Abstract: This paper deals with the optimization of the fuel consumption for a city bus of the city of Rouen, in France. This work takes part from the ANGO project, a french PREDIT-ANR project. The aim consists in modelling the bus and its fuel consumption in order to formulate a problem of optimmization of the consumption (criterium definition, constraints, initialized variables, ...). Some simulation results are shown under an advisory system to the bus driver and experimental works are presented.
Return to index

We04.02: Semi-active ride control of human seated model and robustness analysis.
Click here for full paper
Authors: Dr Georgios Tsampardoukas     Dr Alexandros Mouzakitis     Miss Foteini Tsampardouka    
Abstract: Abstract: The objective of this paper is to synthesize a novel hybrid semi-active control algorithm as well as to compare the semi-active relative to conventional passive system in terms of human body unweighted RMS acceleration values. A theoretical model of the human seated model is developed in order to simulate the vertical motion of the truck driver. The seated human model is attached on the truck seat model and semi-active control is applied between the excitation base and the moving mass of the truck seat. Algorithm robustness to parametric variations as well as to real-life implementation issues such as feedback signals noise are investigated as well. The results indicate that the injected noise slightly affects the system performance. The vertical acceleration of the human body is significantly reduced using the novel hybrid control algorithm relative to passive system. Hence, the human comfort due to vertical vibrations is substantially increased. Similar results are observed when random excitation (using spectral densities) is employed.
Return to index

We04.03: Asymptotic Tracking applied to the Control of a Turbocharged Diesel Engine
Click here for full paper
Authors: Mr Marcelin Dabo     Dr Nicolas Langlois     Pr Houcine Chafouk    
Abstract: In this paper we propose to present Asymptotic Tracking applied to the tracking problem for a Turbocharged Diesel Engine (TDE). Our goal is to track desired values of TDE which are the gas pressure in the intake manifold and the compressor mass flow rate. Nevertheless for this system with its chosen outputs one faces to the known problem of non-minimum phase systems. To avoid this, the problem of tracking of desired values of the original output y is replaced by that of tracking a suitable constructed modified output ~y for which the values to be tracked are specifically chosen: namely, when the modified output approaches them, the original output converges to the desired values. Simulation results are presented to highlight efficiency of the controller.
Return to index

We04.04: Constrained Variance Control of Peak Pressure Position by Spark Ionization Feedback
Click here for full paper
Authors: Dr Andrew Shenton     Mr Nicholas Rivara     Dr Paul Dickinson    
Abstract: A neural-network (NN) based scheme is presented for control of cylinder peak pressure position (PPP) by spark ignition (SI) timing in a gasoline internal combustion (IC) engine. Spark-ionization current from the spark plug is used to act as a virtual PPP sensor. Off-line training using principal component analysis (PCA) data predicts the cylinder peak pressure position under varying engine load, speed and spark advance (SA) settings. Results demonstrate that the PPP prediction of the NN correlates well with those measured from in-cylinder pressure sensors. A constrained-variance (CV) technique, which is a robustified form of minimum-variance (MV) controller, is designed and applied to regulate the PPP by SA control action. This is validated by experimental implementation on a port fuel-injected (PFI) 4-cylinder 1.6l gasoline internal combustion (IC) engine.
Return to index

We05. Hybrid Vehicles (invited)

We05.01: Modelling and Control of a novel SOFC-IC Engine Hybrid System
Click here for full paper
Authors: Mr Alexandros Plianos     Miss Anita Chaudhari     Prof Richard Stobart    
Abstract: A novel configuration of a Solid Oxide Fuel Cell-Internal Combustion (SOFC-IC) engine is presented and a nonlinear dynamic model that captures the transient phenomena of this system is developed. A variable geometry turbocharger and a throttle present at the air inlet is used to regulate the interacting flows in the combined system. A controller is developed to regulate the output to demand specific setpoints which correspond to the required power output of the hybrid vehicle. The controller is derived by means of identified linear models. It consists of a feedback term, an integral term and a feedforward term. An observer is used for the estimation of the system states. The nonlinear system is assessed under closed-loop control with simulations.
Return to index

We05.02: ELECTRICAL ARCHITECTURES FOR HYBRID VEHICLES: IMPLICATIONS FOR MODELLING AND CONTROL
Click here for full paper
Authors: Dr James Marco    
Abstract: Contained within this paper is a discussion into the modelling and control of the electrical architecture for a HEV. Two configurations of electrical architecture are discussed; a system in which the bus voltage is allowed to vary during vehicle acceleration and regenerative braking and secondly, a fixed bus voltage system in which the voltage is held constant by the inclusion of a bi-directional DC-DC converter. The relative merits of each solution are discussed. Consideration is given to the component sizing of the energy storage device, the associated control system complexity and finally the performance of the HEV
Return to index

We05.03: A highly modular simulation model for hybrid electric fuel cell power drive trains
Click here for full paper
Authors: Dr Volker Pickert     Mr Steve Naylor    
Abstract: Recent computer simulation has demonstrated that there are currently no hybrid electric fuel cell power drive train topologies that give both high performance and high efficiency [1], [2]. In [2] more then ten different hybrid power drive trains have been studied and compared based on simulation. This paper describes for the first time the simulation tool and control algorithms used. The challenge was to design a single simulation model for power drive trains which could simulate more then 10 completely different power drive trains within a limited time period. Prudent design sense suggested that instead of designing a new simulation model for each topology, designing a library of modular components and then creating each topology from these components would be the most efficient way of creating the simulation models. Not only would it ensure that common components are the same across all topologies, but also that if any changes needed to be made to any part of the model the change could be implemented across all topologies with speed and ease. Maintaining up to 10 different models for each of the topologies would quickly have become unmanageable, hard to document and cast doubt on the comparability and reliability of the simulation data. To enable the modular design, interconnections between the different components of the drive trains were defined and standardised to ensure that a single architecture control system could control each and every topology with only minor changes of control system parameters. The control system was designed using a single control loop to ensure that the simulations could be run on a standard desktop computer in an acceptable time. The simulation software Matlab/Simulink was chosen for the study published in [1] and [2] because it is a proven, industry standard package. The paper will discuss in detail the models, standardisation of parameters and the implemented control algorithm. It will discuss the weaknesses and advantages of the developed simulation tool and will address common design errors. [1]Pickert V, Naylor S. Overview of Power Drive Trains for Hybrid Fuel Cell Electric Vehicles. EET-2007 European Ele-Drive Conference Brussels, Belgium, May 30 – June 01, 2007. [2] Pickert V, Naylor S. A review of power drive trains for hybrid fuel cell electric vehicles (HFCEV). 3rd IET Conference on Automotive Electronics, Warwick, 28-29 June 2007.
Return to index

We05.04: NEDC Based Compensated Forward Simulation Approach with Energy Management for Parallel Hybrid Electric Vehicles
Click here for full paper
Authors: Dr John Economou     Mr Piranavan Suntharalingam     Prof Kevin Knowles    
Abstract: This paper presents the power management strategy for a parallel hybrid electric vehicle (PHEV). The vehicle is powered by dual energy sources consisting of internal combustion engine (IC engine) as the peak power source and the electric battery as the secondary energy source. The operational principle of the power management strategy and the possible power flow patterns are described. Based on the power flow and energy availability of the energy sources, the decision-making and the relevant switching function are designed to facilitate the effective power sharing between the two sources. The standard NEDC velocity profile has been utilized via a suitable compensator that feed the data into the forward mathematical model. The obtained results indicated the strengths of such a hybrid topology.
Return to index

We05.05: Block-Control Methods for Low-Order Automotive Control
Click here for full paper
Authors: Dr Andrew Shenton     Mr Christopher Ward    
Abstract: Robust linear and nonlinear control is a continuing requirement for automotive powertrain controls. Newton iteration techniques have been proposed for both nonparametric linear and recently nonlinear control. Such nonparametric methods may eventually allow benefits of both low-order controllers and more rapid calibration time. This paper evaluates the feasibility of such Newton iteration techniques by an experimental comparison of a standard Riccati method, a Riccati J-spectral factorisation and a novel $l_{2}$ algebraic J-spectral factorisation using Newton iteration techniques in a SI engine idle controller. The methods are each applied in a 2-block $H_{\infty}$formulation. The results of experimentally implementing robust idle speed controllers show broadly similar outcomes for all the methods compared and thus indicate the potential of the Newton iteration methods for further development in more advanced nonparametric, low-order and nonlinear control.
Return to index

We06. Robotics: Vision and tracking

We06.01: Visual Tracking System for the Welding of Narrow Butt Seams in Container Manufacture
Click here for full paper
Authors: Mr Zhiguo Yan     Mr De Xu    
Abstract: In this paper, a vision based seam tracking system is proposed for the butt welding in the container manufacture. First, the system structure is designed. Then, the main parts of the system are discussed. The system’s working principle is analyzed. And the algorithms especially the image processing algorithm and control algorithm are proposed. Finally, experiments are conducted to demonstrate the effectiveness of the designed system and the proposed method.
Return to index

We06.02: Homing, Calibration and Model-Based Predictive Control for Planar Parallel Robots
Click here for full paper
Authors: Kvetoslav Belda     Pavel Pisa    
Abstract: Parallel robots represent way to considerably improve accuracy and speed of industrial machine tools and their centres. This paper deals with the preparatory operations: homing and calibration, which precede start-up of the robot work, i.e. real control process. Their procedures are discussed with respect to planar parallel robots and their control. In this paper, as a control strategy, the model-based predictive control is considered. The predictive control offers operator to continously influence the control process. The control issues of planar parallel robots are discussed here.
Return to index

We06.03: Visual servoing control for line and object detection and following using a robotic arm manipulator mounted real time camera system
Click here for full paper
Authors: Dr James Buckle     Mr Andrew Tickle     Dr Fan Wu     Mr Paul Harvey     Prof Jeremy Smith    
Abstract: This paper presents an investigation into compliant robotic systems, focusing on the use of high speed visual servoing without modelling to correct manipulator elasticity when working either above the manufacturers recommended velocities or load levels. This work is based on a PUMA 500 series manipulator in SLAVE control, interfaced with a PC running RTAI Real-time Linux and a PixeLINK CMOS camera. This paper focuses on the details of how the system was designed from the special compliant link to mount the camera, with the associated resonance calculations, to the testing mechanism which consists of trying to get the robot to find the centre of a target line and how it responds to target step-change input. The control response graphs and the accuracy of the system will be discussed in detail. Also included here is how the same system could be implemented using Altera's DSP Builder graphical block interface that could see the systems control mechanism improve in performance and cost.
Return to index

We06.04: Walking Control Algorithm based on Polynomial Trajectory Generation
Click here for full paper
Authors: Dr. Marco Perez-Cisneros     Dr. Erik Cuevas-Jimenez     Dr. Daniel Zaldivar-Navarro    
Abstract: Humanoid walking trajectory is a complicated task because of the high number of degrees of freedom (DOF) and the variable mechanical structure during walking. A non-trivial problem in bipedal robot walking is the instability produced by violent transitions between different walk phases. This work presents a trajectory generation algorithm for a biped robot. The algorithm is based on cubic Hermitian polynomial interpolation of the initial conditions of the robot. This guarantees a smooth transition in the walking phases reducing significantly the tendency for falling down when the walking speed increases or the terrain conditions changes. The algorithm was successfully tested on the biped robot "Dany walker", which was designed at the Freie Universitat Berlin, Germany and the University of Guadalajara, México.
Return to index

We06.05: Experimental Evaluation of Haptic Control for Human Activated Command Devices
Click here for full paper
Authors: Mr. Andrew Zammit Mangion     Dr. Simon G. Fabri    
Abstract: Haptics refers to a widespread area of research that focuses on the interaction between humans and machine interfaces as applied to the sense of touch. A haptic interface is designed to increase the realism of tactile and kinesthetic sensations in applications such as virtual reality, teleoperation, and other scenarios where situational awareness is considered important, if not vital. This paper investigates the use of electric actuators and non-linear algorithms to provide force feedback to an input command device for providing haptics to the human operator. In particular, this work involves the study and implementation of a special case of feedback linearization known as inverse dynamics control and several outer loop impedance control topologies. It also investigates the issues concerned with force sensing and the application of model based controller functions in order to vary the desired inertia and the desired mass matrix. Results of the controllers’ abilities to display any desired impedance and provide the required kinesthetic constraint of virtual environments are shown on two experimental test rigs designed for this purpose.
Return to index

We06.06: Path Planning Generation in Mobile Robots using Evolutionary Harmonic Potential Field Technique
Click here for full paper
Authors: Dr Luis Gonzalez     MC Roberto Reyes    
Abstract: This paper describes a path planning technique for mobile robotics working on cluttered environments. By means of sensor recognition and with a technique of evolutionary harmonic potential fields, the general path planning is reduced to the union of local paths derived by using an optimal genetic algorithm
Return to index

We07. Control Applications : Aerospace

We07.01: Robust, Power Aware Mobile Agent Tracking using an 802.15.4 Wireless Sensor Network.
Click here for full paper
Authors: Mr. Michael Walsh     Dr. Martin Hayes    
Abstract: This work presents an experimental analysis for a lean power, 802.15.4 wireless sensor network based mobile agent tracking problem. A localization procedure is designed that robustly tracks a moving agent despite significant uncertainty existing on the received signal strength vector. The benefits of dynamic power control are considered at two separate levels within the network topology. Firstly, active management of the uplink connection between the stationary tracking reference nodes and a base station is critically assessed. The cost performance benefit that arises from the use of additional feedback bandwidth, where available, and also the design of effective time delay compensation is discussed within this paradigm. Secondly, an additional power control loop is presented where the effects of Raleigh fading and varying time delay on the uplink between mobile node and base station are major factors influencing system performance.
Return to index

We07.02: Real-time trajectory generation technique for dynamic soaring UAVs
Click here for full paper
Authors: Mr Naseem Akhtar     Dr James Whidborne     Dr Alastair Cooke    
Abstract: This paper addresses the problem of generating real time trajectories for the dynamic soaring of UAVs (unmanned aerial vehicles). The aircraft soar using the wind shear available over the oceans. The UAVs utilize the energy from low-altitude wind gradients to reduce fuel consumption. For a propeller driven UAV, a performance index is selected to minimize the average power required per cycle. The control problem is formulated by considering the equations of motion, operational constraints, initial conditions and terminal conditions that enforce a periodic flight. The differential flatness property of the equations of motion are used to transform the problem to the output space, which permits rapid solution using standard nonlinear programming. The results obtained are compared with those achieved for a collocation technique and a constrained optimization technique.
Return to index

We07.03: A Lateral Directional Flight Control System for the MOB Blended Wing Body Planform
Click here for full paper
Authors: Mr Naveed Rahman     Dr James Whidborne    
Abstract: In this work, analysis and design of a lateral directional flight control system for a Blended Wing Body (BWB) aircraft is considered. The BWB configuration chosen for this purpose is the Europeon MOB (Multidisciplinary Optimization Blended Wing Body) planform. The MOB configuration does not have vertical control surfaces for directional stability, instead small winglets with rudders are used. The lateral directional behavior of the baseline MOB configuration is analyzed and the inherent deficiencies both in terms of directional stability and control power are highlited. A modification to the MOB BWB configuration is then proposed in which two vertical rudders are placed at the trailing edge of the center body. The improvement gained in the stability and directional stiffness is then compared with the baseline configuration. The open loop analysis is then followed by the design of a yaw damper for both configurations. It is concluded that the baseline MOB configuration with winglet rudders does not have enough trim/control authority especially under asymmetric thrust conditions. Rudder control has to be modified in order to make this a practical design.
Return to index

We07.04: Suppressing aeroelastic vibrations via stability region maximization and numerical continuation techniques
Click here for full paper
Authors: Dr Max Demenkov     Dr Mikhail Goman    
Abstract: An active flutter suppression using linear sub-optimal control scheme is investigated for a 2dof airfoil system with nonlinear torsional stiffness and limited deflection amplitude of its single actuator. The suppression of limit cycle oscillations in the nonlinear closed-loop system is achieved through maximization of the stability region of its linearized system. The critical value of the control input amplitude is determined via numerical continuation of closed-loop limit cycle. At this value, the cycle experiences saddle-node bifurcation and disappears, satisfying the necessary condition for the global stability in the closed-loop system.
Return to index

We07.05: ANFIS Network Design Method for Modelling of the Twin Rotor MIMO System (TRMS)
Click here for full paper
Authors: Mrs. Siti Fauziah Toha     Dr. M. O. Tokhi     Mr. Zakaria Hussain    
Abstract: Interest in system identification especially in the field of nonlinear system has alarmingly increased in the past few decades. Soft-computing methods which concern computation in an imprecise environment have gained significant attention amid widening studies of explicit mathematical modelling. In this research, adaptive neurofuzzy inference systems (ANFIS) network design is deployed and used for modelling a Twin Rotor MIMO system. The system is perceived as a challenging engineering problem due to its high nonlinearity, cross coupling between two axes and inaccessibility of some of its states and outputs for measurements. Accurate modelling of the system is thus required to be developed to achieve control objectives satisfactorily. It is demonstrated experimentally that ANFIS can be effectively used as a mean of nonparametric modelling with a highly accurate result. Model validation tests including training and test validation and correlation tests were finally carried out in order to validate the model. Keywords: Soft-computing, Adaptive Neurofuzzy Inference Systems, TRMS, Nonparametric modelling
Return to index

We07.06: The optimisation of stator vane settings in multi-stage axial compressors using a particle swarm optimisation
Click here for full paper
Authors: Mr Hongsuk Roh     Prof. Steve Daley    
Abstract: Abstract ─ Axial flow compressors are required to operate over a wide range of mass flow rate and rotational speeds at high efficiency in industrial gas turbines. However, the useful range of operation of the axial compressor is limited by the onset of two instabilities known as surge and rotating stall. To resolve these problems, variable stator blades or VGVs are considered by optimising the blade setting in order to avoid the stall and subsequent surge. To investigate performance, particularly obtaining acceptable convergence time for practical purposes, a steady state model of a 15 stage multi-axial compressor is utilised. For the effective search for an optimum setting, the variation in VGVs with respect to a different combination of objective functions is considered. In this paper, a particle swarm optimisation method with time-varying inertia weight factor was proposed and utilised to obtain the best value for a normalised objective function. The results of PSO demonstrate the effectiveness and the suitability of its use in this proposed application. Index Terms ─ axial compressor, particle swarm optimisation, variable guide vanes (VGVs)
Return to index

Th01. Control Theory: Discrete systems

Th01.01: IMPLEMENTATION OF NON-UNIFORM SAMPLING FOR ‘ALIAS-FREE PROCESSING’ IN DIGITAL CONTROL
Click here for full paper
Authors: Mr Mohammad Samir Khan     Prof Roger Goodall     Dr Roger Dixon    
Abstract: A non-uniform additive pseudo-random sampling pattern (mainly proposed in the signal processing communities) can be used for performing an ‘alias-free signal sampling’ process. The carefully designed sampling scheme can mitigate the effects of aliasing and permit significant reductions in the average sampling frequency, leading to more efficient processor utilization. Despite the fact that the sampling scheme potentially yields a number of advantages, has previously received no significant attention in the field of Control theory for research. This paper highlights the implementation of this technique in digital control compensators, discussing the importance of selecting a suitable form for implementation and illustrates the potential benefits in terms of alias avoidance.
Return to index

Th01.02: Extraproximal Method for Markov Chains Finite Games
Click here for full paper
Authors: M.S. Samuel Moya     P.hD. Alexander Poznyak    
Abstract: In this paper a regularized version of the"extraproximal method" is suggested to be applied for finding a Nash equilibrium in a multi-participant finite game where the dynamics of each player is governed by a finite controllable Markov chain. The suggested iterative technique realizes the application of a two-step procedure at each iteration: at the first (or preliminary) step some "predictive approximation" of the a current approximation is calculated; at the second step (the main step of the iteration) this prediction is used to complete the current iteration. The convergence of the suggested procedure to one of Nash-equilibrium is analyzed. The conditions guaranteeing this convergence are discussed. The numerical example demonstrates a good workability of the proposed approach.
Return to index

Th01.03: $L_2 $ gain analysis for linear discrete switched delay systems
Click here for full paper
Authors: Dr. Xi-Ming Sun    
Abstract: The problem of $L_2$ gain for a class of linear discrete switched systems with disturbance input is considered in this paper. It is assumed that not all of the subsystems have normal $L_2$-gain. Based on the average dwell time method, we search for switching signals to make the switched systems achieve the weighted $L_2$ gain. As a special case, the criterion of normal $L_2$ gain under arbitrary switching is also developed. Without considering delays, the proposed result degenerates to existing one.
Return to index

Th01.04: Development of second order plus time delay (SOPTD) model from orthonormal basis filter (OBF) model
Click here for full paper
Authors: Dr Ramasamy M     Mr. Lemma D Tufa     Prof Patwardhan Sachin C     Dr Shuhaimi M    
Abstract: A novel method to determine the parameters of a second order plus time delay (SOPTD) model from a step response is presented. The method is uniquely effective in developing SOPTD models from Orthonormal Basis Filter (OBF) model. A noise free OBF model can be easily developed from a noisy response data and any type of input with a crude estimate of time constants and no-prior knowledge of time delay. The OBF model developed in this manner can capture the dynamics of a process with only a few numbers of terms (parsimonious in parameters) and do not have the problem of inconsistency which is commonly encountered in ARX models. In addition, the OBF model gives the liberty to use any type of input sequence for identification so that we can design the best possible input sequence. However, the time delay in OBF models is estimated by a non-minimum phase zero and current methods of developing SOPTD model from a step response cannot be applied effectively. In this paper, an effective method to identify SOPTD systems or for approximating higher order systems by SOPTD mode from OBF models is proposed. The efficacy of the proposed method is demonstrated through simulation studies.
Return to index

Th01.05: Improved FOPDT model estimation with Delayed-relay feedback for constant time dominant processes
Click here for full paper
Authors: Ms. Zeinab Tehrani Zamani     Mr. Behzad Moshiri     Mr. Ali Khaki Sedigh     Mr. Alireza Fatehi    
Abstract: In this paper with reference to analytical results of different well-known relay feedback methods, we illustrate a main deficiency in parameter estimation of processes with a small ratio of time delay to time constant. Then to rectify this problem we introduce a modified relay feedback structure with additional delay to estimate the parameters of the FOPDT transfer function of the system. The significance of this method lies in the fact that many industrial plants perform fairly such as FOPDT systems, and a wide range of processes have negligible dead time versus their long constant time. Also, the estimated FOPDT transfer function from proposed relay feedback test can be used as a priori knowledge in advanced control strategies which need a FOPDT model of the system. The method is straightforward and simulation results illustrate the effectiveness, and simplicity of the proposed method.
Return to index

Th01.06: Reduced-order Local Optimal Controller for a Higher Order System
Click here for full paper
Authors: Miss Zukhra Kamalova     Mr Mahmoud Ashry     Dr Tim Breikin    
Abstract: In this paper, a reduced order local optimal controller is designed for a higher order system. A reduced order model is obtained for the higher order system and its parameters are used for the reduced order local optimal controller. Also, genetic algorithm is used with the reduced order local optimal controller structure to design the controller parameters instead of obtaining them from the reduced order model. These results obtained are compared with the results obtained from full order local optimal controller. Finally, analogy between reduced order local optimal controller and PI controller parameters is represented. As such, this reduced order local optimal controller can be used for tuning PI controller parameters. Experimental results on a lab-based test rig confirm the effectiveness of the reduced order local optimal controller.
Return to index

Th02. Robotics: Control and recognition

Th02.01: EXPERIMENTAL STUDIES OF MULTI-ROBOT FORMATION AND TRANSFORMING
Click here for full paper
Authors: Mr. Lei Liu     Prof. Yongji Wang     Prof. Shuanghua Yang     Mr. Graham Watson     Mr. Brian Ford    
Abstract: Focusing on multi-robot flocking, this paper develops a formation holding and transforming method by using the leader-follower strategy on the double or triple robot groups. With this method the state and role of each robot can be identified, and the most suitable topology of the multi-robot formation is decided in order to go through a gap or avoid an obstacle. Furthermore, the algorithms and strategies are implemented on the Koala Robots of SEIC/BAe systems, and the communication between each robot is based on the Internal Communication Engine (ICE), which is a popular middleware used for building distributed communication environments. Finally, the experimental results demonstrate the efficiency of the proposed method.
Return to index

Th02.02: Control Laws Design and Simulation Validation of Autonomous Mobile Robot Off-Road Trajectory Tracking
Click here for full paper
Authors: Dr. Yang Yi     Pr. Fu Mengyin     Pr. Sun Changsheng     Pr. Wang Meiling    
Abstract: Abstract: Autonomous wheeled skid-steering mobile robot off-road trajectory tracking is focused on. According to the kinematics and dynamics analysis of the robot, a constraint of the robot motion is put forward. As uncertain disturbance factors exist during the robot off-road running, a novel fuzzy lateral control law is proposed, which makes the robot motion globally asymptotically stable. According to the requirement of off-road running, the longitudinal control law and the sensor pan-tilt control law are also presented. Based on ATRV2 mobile robot and the off-road terrain information, using virtual prototype technology, ADAMS and MATLAB co-simulation platform is established, and the robot simulation running experiment, the off-road trajectory tracking, is performed in the environment. The simulation results indicate that the control laws are robust and effective for the mobile robot off-road running.
Return to index

Th02.03: Fast Gabor Filters for Object Recognition of Mobile Robot
Click here for full paper
Authors: Xiaorong Wang     Yingkai Zhao     Jinguo Lin    
Abstract: Gabor filters have been used extensively in areas related to feature extraction of images due to their localization in space and bandlimited properties. Since 2-D Gabor filters have more complexity of computation, they have been used in static image decompositions rather than for mobile robots. In an attempt to reduce complexity of computation of 2-D Gabor filters for mobile robots, in this paper, a method of fast Gabor filters is presented. In the method, 2-D Gabor filters are decomposed into 1-D Gabor filters along non-orthogonal axes with different variances first, and those 1-D Gabor filters are recursively implemented, then the image group after fast Gabor filters is extracted feature by Principle Component Analysis (PCA), last the image would be classified by support vector machine (SVM). Experiment results indicate that, mobile robots can reach recognition rate of more than 92\% and speed of quasi real-time image processing of 8 frames per second by the method.
Return to index

Th02.04: Sub-Optimal Control Based on Passivity for Euler-Lagrange Systems.
Click here for full paper
Authors: MC Jesús P. Ordaz-Oliver     PhD Omar A. Dominguez-Ramirez     MC Filiberto Muñoz-Palacios    
Abstract: This paper, present a class of nonlinear control, the feedback error scheme is proposed for trajectory tracking of an Euler-Lagrange system. The controller in this paper has the advantage of global stability and robustness, moreover, we provide a passivity based on stability analysis which suggest that the system has a condition of strictly semi-definite positive realness of tracking error dynamics, this is a necessary condition for a global stability, to this end the explicit solution of the Hamilton-Jacobi-Bellman principle found by solving the Lyapunov function. In order to demonstrate the control approach, we present a simulation using a 3-DOF robot, to this case we use Phantom Haptic device dynamical model.
Return to index

Th02.05: Control Based on Energy for Vertical 2 Link Underactuated Robots.
Click here for full paper
Authors: MC J. Patricio Ordaz-Oliver     PhD Omar A. Dominguez-Ramirez     MC Eduardo S. Espinoza-Quesada    
Abstract: A class of nonlinear control scheme for swinging up and stabilization of underactuated 2-link robots is introduced. To this end, the control law proposed is applied to an benchmark system. The proposed methodology is designed based on Euler-Lagrange dynamics, energy analysis and Lyapunov theory. A class of linear control doesn't allow to compensate the no linear dynamics performance, for example, inertia, Coriolis, gravity and tribology forces, specially when the system present the underactuated property. The controller in this paper has the advantage of local stability, moreover, we provide a passivity based on stability analysis which suggest that the system has a condition of strictly semi-definite positive realness of tracking energy error and desired position, this is a necessary condition for a local stability. Swinging control is based on an energy approach and the passivity properties, and then some conditions on the parameters in the control law such that the total energy of the underactuated robot converges to the potential energy of its top upright position are given. The stabilization system is based on switching LQR control.
Return to index

Th03. Control Methodology 1

Th03.01: CHEAP COMPUTATION OF OPTIMAL REDUCED MODELS USING SYMBOLIC COMPUTATION
Click here for full paper
Authors: Prof. Oluwafemi Taiwo     Mr. Akinola Falola    
Abstract: Many methods have been proposed for the reduction of single-variable and multi-variable systems having ordinary denominators and those having delays in their numerators and denominators. In this paper, an algorithm for this purpose is proposed that can be easily used by anybody with not too advanced knowledge of mathematics yielding optimal reduced models. This algorithm makes use of the symbolic capabilities of computer algebraic systems (CAS) like Mathematica, MATLAB and Maple to carry out the model reduction. An advantage of the algorithm is that it can be easily automated and used for rational, irrational, retarded SISO and MIMO systems.
Return to index

Th03.02: On the Relative Degrees and the Interactor Matrix of Linear Multivariable Systems
Click here for full paper
Authors: prof Yasuhiko Mutoh    
Abstract: In various types of control problems for linear multivariable systems, an interactor matrix is often used instead of the relative degrees. The interactor is a polynomial matrix of Laplace operator "s" which cancels all zeros at infinity of the system transfer matrix by multiplying from the left. This implies that the interactor is another expression of zeros at infinity of the linear multivariable system and then there should be a direct relation between the structure of the interactor and the structure of zeros at infinity. However, the interactor is not determined uniquely for the given system. This paper shows that the structure (multiplicities) of zeros at infinity of the linear multivariable system coincides with the row degrees of its interactor, if and only if the interactor is row proper.
Return to index

Th03.03: An Approach to Pole Placement Method with Output Feedback
Click here for full paper
Authors: Gra. St Selda GUNEY     As Prof Ayten ATASOY    
Abstract: In this paper, we have presented a simple method in order to solve the pole placement problem of linear output feedback systems with state space models. Pole placement is a control method assigned to arbitrary closed loop poles by state or output feedback. In linear systems, poles have influence on stability, system response, transient response, and band width. Pole placement methods are used in the design of different control systems. This paper presents a numerical algorithm for pole placement with output feedback. Earlier, a method based on Sylvester’s equation had been applied to pole placement with state feedback. In our work, this method is applied to pole placement with output feedback. It has been obtained by using generalised inverse approach. The pseudo-inverse C+ of an m-by-n output matrix C caused a problem while output feedback matrix Ko is calculated. The generalised inverse approach is used for overcoming this problem. It is shown that real different poles and repeated poles are assigned via output feedback by using the given algorithm. The efficiency of the algorithm is denoted with several extensive numerical examples. Also the performance of the method is tested for different poles on various systems. The results are compared with generalised mapping approach.
Return to index

Th03.04: Robust output-feedback tracking control of multivariable continuous-time systems in an LMI setting
Click here for full paper
Authors: Tansel Yucelen    
Abstract: This paper presents a new linear matrix inequalities (LMI) based approach for sub-optimal output-feedback tracking control of continuous-time multiple-input multiple-output (MIMO) systems. The proposed method is robust and capable of tracking any given constant or time-varying references via minimizing the error between these reference signals and the states of the MIMO system with a desired level of attenuation. The approach here is based on $H_{\infty}$performance index for state-feedback control design and H2 performance index for state-observer design. In addition, the proposed methodology has proven to be stable. The contribution of this paper is demonstrated through a detailed multivariable system simulation for the proof of concept.
Return to index

Th03.05: Stabilizing systems with aperiodic sample-and-hold devices: state feedback case
Click here for full paper
Authors: AProf. Hisaya Fujioka     Mr. Toshiharu Nakai    
Abstract: Motivated by the widespread use of networked and/or embedded control systems, an algorithm for stabilizing sampled-data feedback control systems with uncertainly time-varying sampling intervals is proposed, where it is assumed that the sampled state is available for feedback. The algorithm is an extension of that for stability analysis in the authors' previous study, and is based on the robustness against the variation of sampling intervals derived by the small-gain condition. The validity of the algorithm is demonstrated by numerical examples.
Return to index

Th03.06: ROBUST CONTROLLER TUNING BASED ON COEFFICIENT DIAGRAM METHOD
Click here for full paper
Authors: Dr. Mehmet Turan Soylemez     Mr. Omur Ocal     Prof. Atilla Bir    
Abstract: In this paper, Coefficient Diagram Method (CDM), which is a controller design method that provides remarkable time-domain characteristics, is combined with a PI controller in order to design robust controllers. In particular, it has been demonstrated that it is possible to provide robust tuning rules for first order plus time delay (FOPTD) systems. Here, PI controller is used for improving the steady-state response of the system and for providing an extra parameter for tuning robustness. Pole Colouring method is used for measuring robustness. Calculation of robust tuning rules is computationally expensive, since it is required to find the best values of the free parameter of the PI controller for different plants. However, after using a curve fitting algorithm it is possible to obtain simple tuning rules to determine robust controllers.
Return to index

Th04. Condition monitoring and fault diagnosis

Th04.00: Fault Detection for Vehicle Suspensions Based on System Dynamic Interactions
Click here for full paper
Authors: Mr Xuejun Ding     Dr Tianxiang Mei    
Abstract: This paper presents a novel method for the fault detection and isolation for rail vehicle suspensions that explores the additional dynamic interactions between different motions of a bogie or body caused by the failure of suspension components by taking advantage of symmetrical mechanical configurations of railway bogies. The study is focused on the monitoring of the vertical primary suspensions of a conventional bogie vehicle to demonstrate the general principle and effectiveness of the proposed method in detecting damper faults, although the technique is equally applicable for suspensions in other directions.
Return to index

Th04.01: Observer-Based Residual Design for Nonlinear Systems
Click here for full paper
Authors: Mrs Tabassom Sedighi     Dr Ali J. Koshkouei     Prof Keith J. Burnham    
Abstract: This paper presents a method for designing a full order observer for a class of nonlinear system with unknown input in which the nonlinear functions satisfy Lipschitz conditions. The problem of detecting and isolating faults for this class of nonlinear systems are considered and the theoretical results are applied to a mass-spring-damper system in the presence of external disturbances and uncertainties to diagnose the sensor faults.
Return to index

Th04.02: Nonlinear PCA for Transient Monitoring of an Automotive Engine
Click here for full paper
Authors: Prof George Irwin     Dr Xun Wang     Dr Geoff McCullough     Dr Neil McDowell     Dr Uwe Kruger    
Abstract: This paper reports on the application of non-linear principal component analysis to the detection of faults in an automotive gasoline engine during transient operation. An auto-associative neural network is trained on experimental data recorded from an identification cycle in which the engine speed and throttle position inputs were varied over a wide range of the operating map at rates similar to those experienced during normal operation. The model shows good generalisation to the New European Drive Cycle, an absence of unwanted false alarms under fault-free engine conditions, and successful detection of air leaks of varying magnitude in the inlet manifold.
Return to index

Th04.03: APPLICATION OF A PCA MODEL APPROACH FOR MISFIRE MONITORING
Click here for full paper
Authors: Dr Paul King     Prof Keith Burnham    
Abstract: During the calibration and development of a number of on board diagnostics there are several situations which require the analysis of large amounts of multi-dimensional data. One problem that often arises is the comparison of sets of data in order to determine whether there is a significant difference between either two sets of tests or a difference in the results as a consequence of a change in a component. In this paper we investigate this type of problem and develop a PCA model approach to help make such a decision. Data recorded from the validation work on a misfire monitor is used to develop our approach.
Return to index

Th04.04: Robust Fault Isolation for Autonomous Coordination in NCS
Click here for full paper
Authors: Prof. Ron Patton     Mr. Supat Klinkhieo    
Abstract: A recent study shows that a given hierarchical decentralized control system architecture may be suitable for autonomous coordination of fault-tolerant control (FTC) in a network of distributed and inter-connected subsystems. This paper focuses on the development of a robust Fault Detection and Isolation (FDI) strategy for this Network Control System (NCS) FTC problem. By using a robust form of the Unknown Input Observer (UIO), the subsystems can be effectively decoupled from each other for diagnostic purposes. The effects of subsystem interactions are removed from the FDI residuals, thus facilitating a powerful way to achieve robust local subsystem FDI. This subsystem isolation forms a part of the decision-making process of the autonomous system coordinator, facilitating a strategy for autonomy in FTC for NCS.
Return to index

Th05. Parameter estimation and data analysis

Th05.01: multivariate statistical analysis of spectroscopic data
Click here for full paper
Authors: Dr. Ognjen Marjanovic     Mr. Haisheng Lin     Prof. Barry Lennox    
Abstract: This paper focuses on the application and comparison of Principal Component Analysis (PCA) and Independent Component Analysis (ICA) using two generic artificially created datasets. PCA and ICA are assessed in terms of their abilities to infer reference spectra and relative concentrations of the constituent compounds present in the analysed samples. The results show that ICA outperforms PCA and is able to identify the reference spectra of all the constituent compounds, while PCA fails to identify one of the constituent compounds in the case of both data sets. Also, ICA estimates relative concentrations of the constituent compounds more accurately than PCA does.
Return to index

Th05.02: Novel algorithms based on conjunction of the Frisch scheme and extended compensated least squares
Click here for full paper
Authors: Mr Tomasz Larkowski     Mr Jens Linden     Dr Benoit Vinsonneau     Prof Keith Burnham    
Abstract: The paper presents a general framework for the Frisch scheme (FS) and the extended compensated least squares (ECLS) technique within which two new algorithms for the identification of single-input single-output linear time-invariant errors-in-variables (EIV) models are proposed. The first algorithm is essentially the FS using a novel model selection criterion. The second method is a modification of the ECLS technique, which utilizes not only the set of overdetermined normal equations, but also the Frisch equation to solve the parameter estimation problem. An extensive Monte-Carlo simulation compares the novel algorithms with existing EIV identification approaches.
Return to index

Th05.03: Parameter Identification for Electromechanical Servo Systems Using a High-gain Observer
Click here for full paper
Authors: Mr. Usama Abou-Zayed     Dr. Zhiwei Gao     Mr. Xuewu Dai     Dr. Tim Breikin    
Abstract: In this paper, a High-gain Observer (HGO)-based identification technique is used to identify the parameters for electromechanical servo systems. The HGO is used for estimating the system states, disturbances due to uncertainty or parameter changing, and output noise. Then, a new model is presented using QR factorization. The estimated observer states show good agreement with the system actual states for noise free and bounded noisy input/output systems. Using model simulations and real-time input-output data gathered from a noisy electromechanical servo system, experimental study is made. It is shown that HGO-based parameters identification has better performance in bounded noise environment compared with the subspace algorithm.
Return to index

Th05.04: Minimum Entropy Parameter Estimation of Bounded Nonlinear Dynamic Systems with Non-Gaussian state and Measurement noise
Click here for full paper
Authors: Mr George Papadopoulos     Dr Martin Brown    
Abstract: Parameter estimation plays an important role in Systems Biology in helping to understand the complex behavior of signal transduction networks. The problem becomes more complex as the inherent stochasticity of the signaling mechanism involves noise components of non-Gaussian nature. In this paper a novel stochastic parameter estimation method has been developed where the entropy of the joint residual PDF is used as a measure of the systems uncertainty. The optimal parameter values are selected as the ones corresponding to a minimal entropy value of the residual. The novelty of this approach lies in that the assumptions for the system involve both state and measurement noise of arbitrary distribution and the method is designed for general multivariable systems. The residual PDF is approximated using well known Kernel Density Estimation methods. The analysis of the method includes application to the RKIP regulated ERK signaling pathway and comparisons are drawn based on the Least Squares solution of the same problem.
Return to index

Th05.05: Condition Monitoring Approaches to Estimating Wheel-Rail Profile
Click here for full paper
Authors: Dr Guy Charles     Prof Roger Goodall     Dr Roger Dixon    
Abstract: The wheel and rail interaction is the main influence on the dynamic response of a rail vehicle. Any changes in the wheel and rail will change the overall response of the vehicle. The condition monitoring challenge is to interpret these changes into useful condition information. This paper presents the results from initial feasibility studies into model-based condition monitoring at the wheel-rail interface applied to estimating the wheel-rail profile estimation. A number of approaches are presented, based around a Kalman Filter method and least squares methods, applied to a linearised simulation model that included a nonlinear conicity function. The function was successfully estimated using a Kalman Filter that included self-updating information about the shape of the conicity function, and by a piecewise cubic least squares approach.
Return to index

Th05.06: Dynamic Model for the LHIfAM Haptic Interface: Friction parameter estimation
Click here for full paper
Authors: Eng Mildred Puerto     PhD Emilio Sanchez    
Abstract: The approach of new control strategies in the field of haptics usually implies the necessity of having a dynamic model for the user and the haptic robot. This paper presents the procedure applied to obtain the analytical dynamic equations and then the parameter estimation for the LHIfAM haptic device that has been totally developed at CEIT. The proposed methodology begins with the analytical equations computed via the Lagrange-Euler algorithm, then the calculations of masses via CAD models and finally the friction parameters have been obtained via the Least Square method. Finally, the paper explains the validation tests carried out on the estimation results.
Return to index

Th06. Control Theory: Uncertain and time varying

Th06.01: Relay feedback based monitoring and autotuning of processes with gain nonlinearity
Click here for full paper
Authors: Ms. Zeinab Tehrani Zamani     Mr. Behzad Moshiri     Mr. Alireza Fatehi     Mr. Ali Khaki Sedigh    
Abstract: Performance assessment and monitoring of control systems can be used to improve the performance of industrial processes. In this paper, a novel relay feedback based method for monitoring and automatic retuning of a class of proportional-integral (PI) controllers is proposed for the systems with gain nonlinearity. For performance assessment of the closed loop system, a time domain evaluation criteria based on the integral of the absolute value of the error (IAE) and the normalized pick of the error in setpoint (SP) changes are presented. Simulation results on the highly nonlinear pH process have shown the effectiveness and feasibility of this method.
Return to index

Th06.02: Design and Implementation of a Time Varying Local Optimal Controller based on RLS Algorithm for Multivariable Systems
Click here for full paper
Authors: Mr Mahmoud Ashry     Mr Usama Abou-Zayed     Dr Tim Breikin    
Abstract: Since the local optimal controller is a model based controller, the controller parameters can be updated with the on-line parameter tuning. Recursive least squared algorithm is used for on-line closed-loop identification of the model parameters. In this paper, the local optimal controller is designed for multivariable system and its parameters are updated on-line. The time varying local optimal controller is implemented on a lab-based test rig. In addition to its computational efficiency and structure simplicity, the experimental results confirm the effectiveness of this controller especially when the parameters of the system are time-variant.
Return to index

Th06.03: CONTROL TECHNIQUES FOR MULTI-AXIS REAL-TIME DYNAMIC SUBSTRUCTURING
Click here for full paper
Authors: Miss Meriem Allouache     Dr. David Wagg     Dr. Mark Lowenberg    
Abstract: Real-time substructuring is a novel hybrid method for the dynamic testing of complex engineering structures. This technique involves creating a hybrid model of the entire structure by combining an experimental test piece with the remainder of the structure which is modelled numerically. The essence of this technique is to emulate the dynamic behaviour of the original structure by using real-time control techniques to join the two substructures together. This paper will focus on control strategies involving delay compensation and synchronization on a multi-axis experimental rig commissioned specifically for real-time substructuring. The current application also inspires studies of coupling in real-time.
Return to index

Th06.04: FDI OF THREE-TANK SYSTEM USING NEUROFUZZY NETWORKS WITH LOCAL APPROACHES
Click here for full paper
Authors: Dr H.T. Mok     Dr C.W. Chan    
Abstract: In this paper, a fault detection and isolation (FDI) scheme is derived based on fuzzy rules extracted from the neurofuzzy network that models the residual of the system. First, a fault database (FDB) is constructed from fuzzy rules extracted from the neurofuzzy networks that model all possible faults in the system. By comparing the currently extracted fuzzy rules with those in the FDB using the nearest neighbour classifier, faults are diagnosed online. As the number of rules in the FDB can be quite large, the FDI scheme proposed here utilises the local approaches to reduce the computation load and to improve the sensitivity of the method. The proposed FDI scheme is successfully applied to diagnose faults in a nonlinear three-tank control system.
Return to index

Th06.05: A foray into P2BL in a Control Systems Course
Click here for full paper
Authors: Mr David Hamilton     Dr Tom O'Mahony    
Abstract: Abstract: Project- and problem-based learning are recommended instructional models that develop students capabilities to solve problems, work in teams and learn independently. This paper presents a course component, developed by the authors, in which the students are presented with an authentic problem and a blended project and problem based learning instructional model is used to develop these transferable skills. The component also integrates international best practice from the field of education. A number of techniques were used to evaluate the course and the results indicate that students perceived that the component developed their ability to work in teams and are very open to similar components being introduced into additional modules. However, the authors noted that students experienced considerable difficulty translating prior knowledge to an unfamiliar scenario and this prompts (as yet unanswered) questions about the effectiveness of traditional teaching models.
Return to index

Th06.06: A NEW APPROACH TO INPUT-OUTPUT PAIRING ANALYSIS FOR UNCERTAIN MULTIVARIABLE PLANTS
Click here for full paper
Authors: Dr Bijan Moaveni     Prof. Ali Khaki Sedigh    
Abstract: In this paper, a new method to analyze the input-output pairing for uncertain multivariable plants is proposed. Here, Hankel Interaction Index Array is used to choose the appropriate input-output pair and a theorem will be presented to show the effect of additive uncertainties on input-output pairing of the system. In this theorem a new approach to compute the variation bound of Hankel Interaction Index Array elements due to additive uncertainties in state space framework is given to study the possible change in input-output pairing. Finally, two typical plants are employed to show the main points of the proposed methodology
Return to index

Th07. Inequality Procedures(invited)

Th07.01: Method of Inequality-Based Multiobjective Genetic Algorithm for Optimizing Cart-Double-Pendulum-System
Click here for full paper
Authors: Prof. Tung-Kuan Liu     Mr. Chiu-Hung Chen     Prof. Zu-Shu Li    
Abstract: This article presents a multiobjective genetic algorithm to tracking the optimal parameterization problem of the controller concerning the swinging-up and handstand-control of the general cart-double-pendulum system (CDPS). The design based on the Human-Simulated Intelligent Control (HSIC) theory is required to meet various criteria according to the expected specifications. The proposed algorithm extends from the original method of inequality-based multiobjective genetic algorithm (MMGA) and can efficiently maintain the Pareto set of the CDPS optimal parameters in the evolutionary population.
Return to index

Th07.02: Design of Critical Control Systems Using Disturbance Cancellation Controllers
Click here for full paper
Authors: Prof. Tadashi Ishihara     Dr. Takahiko Ono    
Abstract: Recently, the authors have proposed a new critical control system design which does not require extensive numerical search. The key idea is to decompose the design problem into two simpler design steps by the technique used in the classical loop transfer recovery method. Since the integral action of the controller is required to deal with the rate-limited exogenous signal, our previous work assumes the use of the Davison type integral controller. In this paper, we discuss the application of the new tmethod to the control system design using the integral controller based on the disturbance cancellation.
Return to index

Th07.03: Development of the actively-controlled beds for ambulances
Click here for full paper
Authors: Dr Takahiko Ono     Dr Tadashi Ishihara     Dr Hikaru Inooka    
Abstract: During transportation by ambulance, the inertial acceleration acts on a patient when an ambulance decelerates or turns a corner. Such acceleration often gives a supine patient physical stress such as blood pressure variation and body sway, which results in pain or a feeling of discomfort. To reduce this undesirable effect of the acceleration, the actively-controlled bed, which controls a posture of the bed to cancel the inertial acceleration by the gravitational acceleration, was developed. This paper gives an overview of its development, including control system design and performance evaluation.
Return to index

Th07.04: Robust Multivariable Control System Design Using The Method Of Inequalities
Click here for full paper
Authors: Prof. Olufemi Taiwo     Mr. Kayode Owa     Mr. Ayodeji Akere     Mr. Temitope Ajetunmobi    
Abstract: This work is concerned with the design of robust control systems for multivariable time-delayed plants using the Method of Inequalities (MOI). It is an extension of previous applications in that the time delays are not approximated by rational functions. A further extension is that the designed systems were required to satisfy certain robustness conditions. This assures guaranteed stability and performance for uncertain systems. Admirable qualities of the method are that it facilitates the design of simple controllers of predetermined structures,such as decentralized controllers,while allowing multi-objective specifications.
Return to index

Th07.05: Poiseuille Flow Controller Design via the Method of Inequalities
Click here for full paper
Authors: Dr James Whidborne     Dr John McKernan     Dr George Papadakis    
Abstract: This paper investigates the use of the Method of Inequalities (MoI) to design output-feedback compensators for the problem of the control of laminar plane Poiseuille flow. In common with many flows, the dynamics of plane Poiseuille flow are very non-normal. Consequently, small perturbations grow rapidly with a large transient that may trigger nonlinearities and lead to turbulence even though such perturbations would, in a linear flow, eventually decay. Such a system can be described as a conditionally linear system. The sensitivity is measured using the maximum transient energy growth, which is widely used in the fluids dynamics community. The paper considers two approaches. In the first, the MoI is used to design low-order proportional and P+D controllers. In the second approach, the MoI is combined with McFarlane and Glover's H-infinity loop-shaping design procedure in a mixed-optimization approach. The results show that the low-order controllers do reduce the maximum transient energy growth but the reduction is not satisfactory. Furthermore, the H-infinity approach does not improve the performance.
Return to index

Th07.06: Computation of Peak Output for Inputs Restricted in $\mathcal{L}_2$ and $\mathcal{L}_\infty$ Norms Using Convex Optimization
Click here for full paper
Authors: Dr Suchin Arunsawatwong     Mr Warit Silpsrikul    
Abstract: Control systems design by the principle of matching gives rise to problems of evaluating the peak output. This paper proposes a practical method for computing the peak output of linear time-invariant and non-anticipative systems for a class of possible sets that are characterized with mixed bounding conditions on the two- and/or the nfinity-norms of the inputs and their derivatives. The associated infinite-dimensional convex optimization problem is approximated as a large-scale convex programme defined in a Euclidean space, which are associated with sparse matrices and thus can be solved efficiently in practice. The numerical results show that the method performs satisfactorily, and that using a possible set with many bounding conditions can help to reduce the design conservatism and thus yields a better match.
Return to index

Th08. Sliding Mode Control

Th08.01: SLIDING MODE CONTROLLERS USING OUTPUT INFORMATION: AN LMI APPROACH
Click here for full paper
Authors: Mr. xiaoran Han     Ms. Emilia Fridman     Ms. Sarah Spurgeon     Mr. Chris Edwards    
Abstract: This paper considers the development of Sliding mode output feedback controllers. The existence problem is solved via a static output feedback formulation using a descriptor approach. Linear matrix inequalities (LMI) are used to obtain the parameters of the switching function. The paper provides conditions in terms of the system structure for a stable reduced-order sliding motion to exist. A controller is constructed to ensure the sliding mode is reached. A numerical example from the literature illustrates the proposed method.
Return to index

Th08.02: Application of MPC and Sliding Mode Control To IFAC Benchmark Models
Click here for full paper
Authors: Dr Meghan McGookin     Dr David Anderson     Dr Euan McGookin    
Abstract: The comparison of Model Predictive Control (MPC) and Sliding Mode Control (SMC) are presented in this paper. This paper investigates the performance of each controller as the navigation system for IFAC benchmark ship models (cargo vessel and oil tanker). In this investigation the navigation system regulates the heading angle of the two types of marine vessel with reference to a desired heading trajectory. In this investigation, the result obtained from MPC is compared with a well-established control methodology, namely Sliding Mode control theory. Wave disturbances and actuator limits are implemented to provide a more realistic evaluation and comparison for the proposed control structure.
Return to index

Th08.03: SLIDING-MODE POSITION CONTROL OF A 1-DOF SET-UP BASED ON PNEUMATIC MUSCLES
Click here for full paper
Authors: Mr Javier Arenas     Mr Aron Pujana-Arrese     Mrs Sandra Riaño     Mrs Ana Martinez-Esnaola     Dr Joseba Landaluze    
Abstract: A one-degree-of-freedom arm driven by pneumatic muscles has been designed and built in order to research the applicability of pneumatic artificial muscles in industrial applications. The experimental set-up is very non-linear and very difficult to control properly. As a reference, an enhanced PID controller was designed. At the same time, a sliding-mode controller based on an observer was designed and implemented. Firstly, this paper presents the experimental set-up and the system’s linear models. After that, it focuses on the process of designing the sliding-mode controller. Finally, some results obtained in simulation as well as experimentally are presented.
Return to index

Th08.04: About Equivalence Between Sliding Mode and Continuous Control Systems
Click here for full paper
Authors: Prof. Ryszard Gessing    
Abstract: It is shown, how to create the continuous system equivalent to the system with sliding mode control. In the case of minimum phase plants, the system arises from the replacement of the relay with small hysteresis by the amplifier with high gain, connected in series with saturation having appropriate parameters. In the case of nonminimum phase (or other difficult plants) it is noted that similar equivalence exists for the continuous and relay system with parallel compensator. The latter system may be treated as the system with modified sliding mode control. In the equvivalent continuous system the chattering effect, related with sliding mode control doesn't exist. \copyright 2008
Return to index

Th08.05: Fuzzy Sliding Mode Controllers for Vehicle Active Suspensions
Click here for full paper
Authors: Dr Ali Koshkouei    
Abstract: In this paper a method using sliding mode control and fuzzy logic techniques is presented for controlling a quarter-vehicle hydraulic active suspension. In this method a set of linear systems is considered which approximately describes the behaviour of the nonlinear suspension model. For each subsystem a suitable sliding mode controller is designed and then based on the Takagi-Sugeno fuzzy method an overall sliding mode controller for the Takagi-Sugeno model is designed. The proposed method considers two phases. In the first phase, the suspension dynamics is controlled via the actuator between the sprung and unsprung masses. Then the spool valve displacement dynamics is considered to control the current of the servo valve. Since there is an unknown parameter in the system an adaptation law is proposed to yield an appropriate estimate.
Return to index

Th08.06: Design of an Asymptotic Sliding Mode Algorithm for Nonlinear Systems: An Observer Based Approach
Click here for full paper
Authors: Prof. Chieh-Li Chen     Mr. Chao-Chung Peng    
Abstract: Using the concept of dynamic sliding mode control, an asymptotic robust controller for high order nonlinear systems is presented. In this approach, a n-order nonlinear system is transformed into a first order system through systematic backstepping design and then it will be shown that the control objective is equivalent to design a robust control law for a 2nd order auxiliary system with only output information available. To get an extra state value, a robust asymptotic observer was integrated into design process that results in an asymptotic sliding mode control algorithm. The proposed method not only preserves some features of conventional sliding mode theory but attenuates undesirable chattering action as well. A numerical example was utilized to demonstrate the applicability of the developed approach.
Return to index

Th09. Imaging and Road Traffic Control

Th09.01: Movement-Based Look-Ahead Traffic-Adaptive Intersection Control
Click here for full paper
Authors: dr.ir. Ronald van Katwijk     prof. Bart De Schutter     prof. Hans Hellendoorn    
Abstract: There exist several control approaches for traffic signal control such as fixed-time, vehicle-actuated, or look-ahead traffic-adaptive control. We argue that in order to flexibly deal with varying demand levels movement-based control (which is already common in vehicle-actuated intersection control) is required instead of stage-based control (which is still employed in the state-of-the-art in look-ahead traffic-adaptive control). The movement-based approach is more flexible than the stage-based approach as it allows green for signals in different stages to start sooner if the demand for all conflicting movements in the current stage has cleared. Therefore, we propose a new movement-based method for look-ahead traffic-adaptive control. The method uses dynamic programming and branch-and-bound algorithms to determine the optimal traffic signal settings. We illustrate via a simulation example that the new approach can significantly outperform vehicle-actuated and stage-based look-ahead traffic-adaptive control.
Return to index

Th09.02: Development of Knowledge-based Measurement with Monocular Vision
Click here for full paper
Authors: Prof. De Xu    
Abstract: In this tutorial, knowledge-based visual measure methods with monocular vision system are investigated. The approaches of visual measure based on knowledge known in advance can be classified to four basic categories according to the types of knowledge, such as point position, line, size or shape knowledge, and motion knowledge. The principle for each category is shortly introduced. Furthermore, visual measure methods based on environment information are also discussed. Finally, conclusion and outlook for the development of knowledge-based visual measure are presented.
Return to index

Th09.03: Multiple Kernel Learning from Sets of Partially Matching Image Features
Click here for full paper
Authors: Dr Siyao Fu     Dr Zengguang Hou     Dr Zize Liang     Dr Min tan     Dr Qi Zuo    
Abstract: Kernel classifiers based on Support Vector Machines (SVM) have achieved state-of-the-art results in several visual classification tasks, however, recent publications and developments based on SVMhave shown that using multiple kernels instead of a single one can enhance interpretability of the decision function and improve classifier performance, which motivates researchers to explore the use of homogeneous model obtained as linear combinations of kernels. Multiple Kernel Learning (MKL) allows the practitioner to get accurate classification results and identify relevant and meaningful features. However, the use of multiple kernels faces the challenge of choosing the kernel weights, and an increased number of parameters that may lead to overfitting. In this paper we show that MKL problem can be formulated as a convex optimization problem, which can be solved efficiently using projected gradient method. Weights on each kernel matrix (level) are included in the standard SVM empirical risk minimization problem with a L2 constraint to encourage sparsity. We demonstrate our algorithm on classification tasks, including object recognition and classification, which is based on a linear combination of histogram intersection kernels, computed at multiple pyramid levels of image encoding, and we show that the proposed method is accurate and significantly more efficient than current approaches.
Return to index

Th09.04: A STUDY ON THE EFFECT OF GPS ACCURACY ON A GPS/INS KALMAN FILTER
Click here for full paper
Authors: Mr King Tin Leung     Dr. James Whidborne     Dr. David Purdy     Dr. Alain Dunoyer     Dr. Robert Williams    
Abstract: In this paper, a Kalman Filter (KF) is used to fuse the Integrated Navigation System (INS) and Global Positioning System (GPS) for the problem of estimating ground vehicle dynamics. Perfect unbiased measurements of the two sensors are extracted from a simulation using IPG CarMaker at a rate of 1 ms to represent a pseudo-analogue signal. Noise is added to the INS and GPS measurements, and then sampled at 100 Hz and 1 Hz respectively. The sampled signals are integrated in the KF and estimated states are compared with the perfect measurements. Results have shown that bias prediction in an INS is achievable using a KF, but highly dependent on the accuracy of GPS. A guided chart is included to aid designers to choose the types of GPS (i.e. sampling rate and variance) against their error criterion.
Return to index

Th09.05: Control Schemes for Safe Operation of Vehicles Convoys
Paper not Submitted
Authors: Dr. Peter Cook    
Abstract: The problem addressed is the selection of control parameters to ensure stable operation of a vehicle convoy system, with maximum throughput, while satisfying constraints imposed by considerations of safety and passenger comfort. Several forms of control law are investigated, including single and multiple look-ahead strategies as well as bi-directional control. By considering simple models, typical conditions for achieving collision avoidance and jerk limitation, while maintaining stability, are derived.
Return to index

Th10. Control Methodology 2

Th10.00: A hands-on approach toward vehicle velocity estimation
Click here for full paper
Authors: Prof. Ansgar Rehm     Hui Fan    
Abstract: Automotive vehicle velocity estimation based only on steering angle and angular wheel velocity measurements is considered in the paper at hand. The approach is based on stationary Kalman filter design combined with a suitable preprocessing of the wheel velocity signals. A detailed assessment of the results by comparison with measured data is given. Possible applications include hierarchical monitoring of vehicle dynamics sensor networks.
Return to index

Th10.01: DATA-DRIVEN DIRECT ADAPTIVE MODEL BASED PREDICTIVE
Click here for full paper
Authors: Mrs Norhaliza Abdul Wahab     Dr Reza Katebi     Dr Jonas Balderud    
Abstract: Abstract: This paper is concerned with the design of Direct Adaptive Model Based Predictive Control (DAMBPC) using subspace identification technique to identify and implement the controller parameters. The direct identification of controller parameters reduces the design effort and computation load which is usually involved with classical adaptive control techniques. The proposed method requires a single QR-decomposition for obtaining controller parameters directly from input-output data when the model dynamic changes. The method using receding horizon approaches to collect data and identify the controller. The paper presents a comparison of performance given by proposed control scheme when applied to a 4-tank nonlinear system with that of a linear model predictive control scheme and multi-loop PID controllers.
Return to index

Th10.02: A New Multi Agent Approach for Traffic Shaping and Buffer Allocation in Routers
Click here for full paper
Authors: Mr Mohamad Taheri Tehrani     Mr Ali Akbar Safavi     Mr Mohamad Raffie Kharazmi     Mr Mohamad Javad Arefi    
Abstract: In this paper, the concepts of reinforcement learning and multi-agent systems are invoked to develop a new traffic shaper for a reasonable utilization of bandwidth while preventing traffic overload in other part of the network. This leads to a reduction in the total number of packet dropping in the whole network. The method is implemented in a novel proposed intelligent simulation environment. The results obtained from this simulation environment show satisfactory behaviors from the aspects of keeping dropping probability low while injecting as many packets as possible into the network in order to utilize the available bandwidth as much as possible. Furthermore, the system can perform well even in situations that have not been previously introduced to the system.
Return to index

Th10.03: Using Lagged Spectral Data in Feedback Control Using Particle Swarm Optimisation
Click here for full paper
Authors: Mr. Caleb Rascon     Prof. Barry Lennox     Dr. Ognjen Marjanovic    
Abstract: The ability to use spectral data within a control loop is beginning to be considered in many areas, particularly in the Pharmaceutical Industry. However, typical spectral analysis tools, such as Classical Least Squares, are very fragile when handling frequency shifts which may occur in spectral measuring devices as a result of poor calibration or external influences. This paper shows that Particle Swarm Optimisation can be used to offset the effect of shift in measured spectra and improve the performance of any control system which may use this measurement.
Return to index

Th10.04: Exact Controls for Superconformal Via Fill Process
Click here for full paper
Authors: Prof. Robert Tenno     Mr. Antti Pohjoranta    
Abstract: This paper reports a means for stabilizing the microvia fill ratio on a desired level, using the total plating time and the system galvanostat setpoint current density as optimal controls. Both control variables are solved as functions of the process state as well as selected manufacturer preference variables that are typical for the via fill technology applied in multilayered printed circuit board production. The optimal controls are obtained as a system of two equations and solved numerically with the gradient descent method. Results of the numerical analysis are presented and discussed.
Return to index