A critique of neural networks for discrete-time linear control

This paper discusses the use of multi-layer perceptron networks for linear or linearizable, adaptive feedback.control schemes in a discrete-time environment. A close look is taken at the model structure selected and the extent of the resulting parametrization. A comparison is made with standard, non-perceptron algorithms, e.g. self-tuning control, and it is shown how gross over-parametrization can occur in the neural network case. Because of the resultant heavy computational burden and poor controller convergence, a strong case is made against the use of neural networks for discrete-time linear control.

A mixture of experts network structure construction algorithm for modelling and control

This paper introduces a new fast, effective and practical model structure construction algorithm for a mixture of experts network system utilising only process data. The algorithm is based on a novel forward constrained regression procedure. Given a full set of the experts as potential model bases, the structure construction algorithm, formed on the forward constrained regression procedure, selects the most significant model base one by one so as to minimise the overall system approximation error at each iteration, while the gate parameters in the mixture of experts network system are accordingly adjusted so as to satisfy the convex constraints required in the derivation of the forward constrained regression procedure. The procedure continues until a proper system model is constructed that utilises some or all of the experts. A pruning algorithm of the consequent mixture of experts network system is also derived to generate an overall parsimonious construction algorithm. Numerical examples are provided to demonstrate the effectiveness of the new algorithms. The mixture of experts network framework can be applied to a wide variety of applications ranging from multiple model controller synthesis to multi-sensor data fusion.

Neurofuzzy control using Kalman filtering state feedback with coloured noise for unknown non-linear processes

This paper presents a controller design scheme for a priori unknown non-linear dynamical processes that are identified via an operating point neurofuzzy system from process data. Based on a neurofuzzy design and model construction algorithm (NeuDec) for a non-linear dynamical process, a neurofuzzy state-space model of controllable form is initially constructed. The control scheme based on closed-loop pole assignment is then utilized to ensure the time invariance and linearization of the state equations so that the system stability can be guaranteed under some mild assumptions, even in the presence of modelling error. The proposed approach requires a known state vector for the application of pole assignment state feedback. For this purpose, a generalized Kalman filtering algorithm with coloured noise is developed on the basis of the neurofuzzy state-space model to obtain an optimal state vector estimation. The derived controller is applied in typical output tracking problems by minimizing the tracking error. Simulation examples are included to demonstrate the operation and effectiveness of the new approach.

A clocking technique with power savings in virtex-based pipelined designs

This paper presents the evaluation in power consumption of a clocking technique for pipelined designs. The technique shows a dynamic power consumption saving of around 30% over a conventional global clocking mechanism. The results were obtained from a series of experiments of a systolic circuit implemented in Virtex-II devices. The conversion from a global-clocked pipelined design to the proposed technique is straightforward, preserving the original datapath design. The savings can be used immediately either as a power reduction benefit or to increase the frequency of operation of a design for the same power consumption.

Pipelining considerations for an FPGA case

This paper presents a semi-synchronous pipeline scheme, here referred as single-pulse pipeline, to the problem of mapping pipelined circuits to a Field Programmable Gate Array (FPGA). Area and timing considerations are given for a general case and later applied to a systolic circuit as illustration. The single-pulse pipeline can manage asynchronous worst-case data completion and it is evaluated against two chosen asynchronous pipelining: a four-phase bundle-data pipeline and a doubly-latched asynchronous pipeline. The semi-synchronous pipeline proposal takes less FPGA area and operates faster than the two selected fully-asynchronous schemes for an FPGA case.

Quantitative evaluation of three reconfiguration strategies on FPGAs: a case study

Reconfigurable computing is becoming an important new alternative for implementing computations. Field programmable gate arrays (FPGAs) are the ideal integrated circuit technology to experiment with the potential benefits of using different strategies of circuit specialization by reconfiguration. The final form of the reconfiguration strategy is often non-trivial to determine. Consequently, in this paper, we examine strategies for reconfiguration and, based on our experience, propose general guidelines for the tradeoffs using an area-time metric called functional density. Three experiments are set up to explore different reconfiguration strategies for FPGAs applied to a systolic implementation of a scalar quantizer used as a case study. Quantitative results for each experiment are given. The regular nature of the example means that the results can be generalized to a wide class of industry-relevant problems based on arrays.

Hardware development of the Yorick series of active vision systems

This paper presents a review of the design and development of the Yorick series of active stereo camera platforms and their integration into real-time closed loop active vision systems, whose applications span surveillance, navigation of autonomously guided vehicles (AGVs), and inspection tasks for teleoperation, including immersive visual telepresence. The mechatronic approach adopted for the design of the first system, including head/eye platform, local controller, vision engine, gaze controller and system integration, proved to be very successful. The design team comprised researchers with experience in parallel computing, robot control, mechanical design and machine vision. The success of the project has generated sufficient interest to sanction a number of revisions of the original head design, including the design of a lightweight compact head for use on a robot arm, and the further development of a robot head to look specifically at increasing visual resolution for visual telepresence. The controller and vision processing engines have also been upgraded, to include the control of robot heads on mobile platforms and control of vergence through tracking of an operator's eye movement. This paper details the hardware development of the different active vision/telepresence systems.

A modular head/eye platform for real-time reactive vision

This paper describes the design, implementation and testing of a high speed controlled stereo “head/eye” platform which facilitates the rapid redirection of gaze in response to visual input. It details the mechanical device, which is based around geared DC motors, and describes hardware aspects of the controller and vision system, which are implemented on a reconfigurable network of general purpose parallel processors. The servo-controller is described in detail and higher level gaze and vision constructs outlined. The paper gives performance figures gained both from mechanical tests on the platform alone, and from closed loop tests on the entire system using visual feedback from a feature detector.

A robust hybrid feedback active noise cancellation headset

This paper investigates the robustness of a hybrid analog/digital feedback active noise cancellation (ANC) headset system. The digital ANC systems with the filtered-x least-mean-square (FXLMS) algorithm require accurate estimation of the secondary path for the stability and convergence of the algorithm. This demands a great challenge for the ANC headset design because the secondary path may fluctuate dramatically such as when the user adjusts the position of the ear-cup. In this paper, we analytically show that adding an analog feedback loop into the digital ANC systems can effectively reduce the plant fluctuation, thus achieving a more robust system. The method for designing the analog controller is highlighted. A practical hybrid analog/digital feedback ANC headset has been built and used to conduct experiments, and the experimental results show that the hybrid headset system is more robust under large plant fluctuation, and has achieved satisfactory noise cancellation for both narrowband and broadband noises.

Performance analysis of the unconstrained FxLMS algorithm for active noise control

The paper analyzes the performance of the unconstrained filtered-x LMS (FxLMS) algorithm for active noise control (ANC), where we remove the constraints on the controller that it must be causal and has finite impulse response. It is shown that the unconstrained FxLMS algorithm always converges to, if stable, the true optimum filter, even if the estimation of the secondary path is not perfect, and its final mean square error is independent of the secondary path. Moreover, we show that the sufficient and necessary stability condition for the feedforward unconstrained FxLMS is that the maximum phase error of the secondary path estimation must be within 90°, which is the only necessary condition for the feedback unconstrained FxLMS. The significance of the analysis on a practical system is also discussed. Finally we show how the obtained results can guide us to design a robust feedback ANC headset.

Enhancing model predictive control using dynamic data reconciliation

The use of data reconciliation techniques can considerably reduce the inaccuracy of process data due to measurement errors. This in turn results in improved control system performance and process knowledge. Dynamic data reconciliation techniques are applied to a model-based predictive control scheme. It is shown through simulations on a chemical reactor system that the overall performance of the model-based predictive controller is enhanced considerably when data reconciliation is applied. The dynamic data reconciliation techniques used include a combined strategy for the simultaneous identification of outliers and systematic bias.

Development and application of a novel algorithm for steady-state optimising control using dynamic information

A novel optimising controller is designed that leads a slow process from a sub-optimal operational condition to the steady-state optimum in a continuous way based on dynamic information. Using standard results from optimisation theory and discrete optimal control, the solution of a steady-state optimisation problem is achieved by solving a receding-horizon optimal control problem which uses derivative and state information from the plant via a shadow model and a state-space identifier. The paper analyzes the steady-state optimality of the procedure, develops algorithms with and without control rate constraints and applies the procedure to a high fidelity simulation study of a distillation column optimisation.

Modelling human dynamics in-situ for rehabilitation and therapy robots

This paper outlines some rehabilitation applications of manipulators and identifies that new approaches demand that the robot make an intimate contact with the user. Design of new generations of manipulators with programmable compliance along with higher level controllers that can set the compliance appropriately for the task, are both feasible propositions. We must thus gain a greater insight into the way in which a person interacts with a machine, particularly given that the interaction may be non-passive. We are primarily interested in the change in wrist and arm dynamics as the person co-contracts his/her muscles. It is observed that this leads to a change in stiffness that can push an actuated interface into a limit cycle. We use both experimental results gathered from a PHANToM haptic interface and a mathematical model to observe this effect. Results are relevant to the fields of rehabilitation and therapy robots, haptic interfaces, and telerobotics

Calibration of closed loop controllers for setting impedances in force-reflecting systems

This paper discusses a new method of impedance control that has been successfully implemented on the master robot of a teleoperation system. The method involves calibrating the robot to quantify the effect of adjustable controller parameters on the impedances along its different axes. The empirical equations relating end-effector impedance to the controller's feedback gains are obtained by performing system identification tests along individual axes of the robot. With these equations, online control of end-effector stiffness and damping is possible without having to monitor joint torques or solving complex algorithms. Hard contact conditions and compliant interfaces have been effectively demonstrated on a telemanipulation test-bed using appropriate combinations of stiffness and damping settings obtained by this method.

Head-controlled assistive telerobot with extended physiological proprioception capability

People with disabilities such as quadriplegia can use mouth-sticks and head-sticks as extension devices to perform desired manipulations. These extensions provide extended proprioception which allows users to directly feel forces and other perceptual cues such as texture present at the tip of the mouth-stick. Such devices are effective for two principle reasons: because of their close contact with the user's tactile and proprioceptive sensing abilities; and because they tend to be lightweight and very stiff, and can thus convey tactile and kinesthetic information with high-bandwidth. Unfortunately, traditional mouth-sticks and head-sticks are limited in workspace and in the mechanical power that can be transferred because of user mobility and strength limitations. We describe an alternative implementation of the head-stick device using the idea of a virtual head-stick: a head-controlled bilateral force-reflecting telerobot. In this system the end-effector of the slave robot moves as if it were at the tip of an imaginary extension of the user's head. The design goal is for the system is to have the same intuitive operation and extended proprioception as a regular mouth-stick effector but with augmentation of workspace volume and mechanical power. The input is through a specially modified six DOF master robot (a PerForceTM hand-controller) whose joints can be back-driven to apply forces at the user's head. The manipulation tasks in the environment are performed by a six degree-of-freedom slave robot (the Zebra-ZEROTM) with a built-in force sensor. We describe the prototype hardware/software implementation of the system, control system design, safety/disability issues, and initial evaluation tasks.