Model predictive control using dynamic integrated system optimisation and parameter estimation (DISOPE)

DISOPE is a technique for solving optimal control problems where there are differences in structure and parameter values between reality and the model employed in the computations. The model reality differences can also allow for deliberate simplification of model characteristics and performance indices in order to facilitate the solution of the optimal control problem. The technique was developed originally in continuous time and later extended to discrete time. The main property of the procedure is that by iterating on appropriately modified model based problems the correct optimal solution is achieved in spite of the model-reality differences. Algorithms have been developed in both continuous and discrete time for a general nonlinear optimal control problem with terminal weighting, bounded controls and terminal constraints. The aim of this paper is to show how the DISOPE technique can aid receding horizon optimal control computation in nonlinear model predictive control.

Behaviour and learning in completely autonomous mobile robots

A number of Intelligent Mobile Robots have been developed at the University of Reading. They are completely autonomous in that no umbilical cord attaches to them to extra power supplies or computer station: further, they are not radio controlled. In this paper, the robots are discussed, in their various forms, and the individual behaviours and characteristics which appear are considered.

Queue-based agent communication architecture for intelligent human-like computer interfaces

Human-like computer interaction systems requires far more than just simple speech input/output. Such a system should communicate with the user verbally, using a conversational style language. It should be aware of its surroundings and use this context for any decisions it makes. As a synthetic character, it should have a computer generated human-like appearance. This, in turn, should be used to convey emotions, expressions and gestures. Finally, and perhaps most important of all, the system should interact with the user in real time, in a fluent and believable manner.

Towards a modular robotics standard for AGV design and programming

Across the world there are many bodies currently involved in researching into the design of autonomous guided vehicles (AGVs). One of the greatest problems at present however, is that much of the research work is being conducted in isolated groups, with the resulting AGVs sensor/control/command systems being almost completely nontransferable to other AGV designs. This paper describes a new modular method for robot design which when applied to AGVs overcomes the above problems. The method is explained here with respect to all forms of robotics but the examples have been specifically chosen to reflect typical AGV systems.

A comparative study of multilayered and single layered neural network based predictive controllers

The authors compare the performance of two types of controllers one based on the multilayered network and the other based on the single layered CMAC network (cerebellar model articulator controller). The neurons (information processing units) in the multi-layered network use Gaussian activation functions. The control scheme which is considered is a predictive control algorithm, along the lines used by Willis et al. (1991), Kambhampati and Warwick (1991). The process selected as a test bed is a continuous stirred tank reactor. The reaction taking place is an irreversible exothermic reaction in a constant volume reactor cooled by a single coolant stream. This reactor is a simplified version of the first tank in the two tank system given by Henson and Seborg (1989).

Trajectory planning of a robot using learning algorithms

The authors consider the problem of a robot manipulator operating in a noisy workspace. The manipulator is required to move from an initial position P(i) to a final position P(f). P(i) is assumed to be completely defined. However, P(f) is obtained by a sensing operation and is assumed to be fixed but unknown. The authors approach to this problem involves the use of three learning algorithms, the discretized linear reward-penalty (DLR-P) automaton, the linear reward-penalty (LR-P) automaton and a nonlinear reinforcement scheme. An automaton is placed at each joint of the robot and by acting as a decision maker, plans the trajectory based on noisy measurements of P(f).

Stable reduced-order models for discrete-time systems

The Routh-stability method is employed to reduce the order of discrete-time system transfer functions. It is shown that the Routh approximant is well suited to reduce both the denominator and the numerator polynomials, although alternative methods, such as PadÃ�Â(c)-Markov approximation, are also used to fit the model numerator coefficients.

A new approach to reduced-order modelling

An error polynomial is defined, the coefficients of which indicate the difference at any instant between a system and a model of lower order approximating the system. It is shown how Markov parameters and time series proportionals of the model can be matched with those of the system by setting error polynomial coefficients to zero. Also discussed is the way in which the error between system and model can be considered as being a filtered form of an error input function specified by means of model parameter selection.

Self-tuning property of state-space self tuners

The basic assumption from implicit self-tuning theory is that, for self tuning to occur, the control input obtained from the estimated system model converges to the value whic would be obtained if the system parameters were known. As as direct result of this, only certain control strategies are acceptable. Here a general rule for the self-tuning property of pole-placement self tuners is obtained, and previous strategies are shown to be special cases of this.