K. C. Cheok

Model reference adaptive control for vehicle active suspension systems

A model reference adaptive control (MRAC) technique for vehicle active suspension subsystems is presented. The MRAC automatically self-tunes the active suspension so that disturbance and vibration of a vehicle is reduced to a level determined by an ideal conceptual suspension reference model. The Lyapunov stability method was used in the design of the MRAC. It is shown that the MRAC suspension can accommodate large variances in sprung load and suspension component characteristics and achieves significant improvements over the passive suspension. Real-time simulation and animation (RTSA) software was developed to provide a visual aid for understanding and interpreting the performance of the MRAC suspension. >

Investigation of Adaptive Control Approaches for Vehicle Active Suspension Systems

A nonlinear Self-Tuning Control (STC) scheme for the nonlinear vehicle active suspension system is investigated and analyzed as an alternative suspension control methodology. The result of the nonlinear STC approach is compared with the linear STC and the model reference adaptive control (MRAC) schemes. The objective of these adaptive control active suspension systems (ACASS) is to tune the controller of the active suspension system automatically such that vehicle disturbance and vibration are reduced to a prespecified level determined by an ideal conceptual suspension system. Thus, an ACASS can accommodate for variances in sprung load and suspension component characteristics. The design and performance analysis of a nonlinear STC suspension system is based on a quarter-car model. Our results show significant improvement in the road handling performance of a high-speed vehicle with the STC active suspension system.

Discrete-time optimal feedback controllers with time-multiplied performance indexes

A class of optimal state and output feedback control laws for discrete-time time-invariant linear systems which minimizes a class of discrete-time time-multiplied performance indexes is presented. A necessary condition, an existence condition, and a sufficient condition for the control laws are derived. A simple example is given to illustrate the effectiveness of the proposed control laws.

Lyapunov stability analysis for self-learning neural model with application to semi-active suspension control system

A supervised training algorithm and an unsupervised learning algorithm for neural models are derived using Lyapunov stability theory and the concept of model reference adaptive control (MRAC). The practicability of the algorithm is demonstrated by its successful application to a semiactive suspension system. The neurocontroller establishes its own control laws for improving the suspension performance without requiring a complete knowledge of the system dynamics. It also makes it possible to evaluate the roles of simple sensors in defining the control laws. Simulation results illustrate the self-improving suspension performance and show that superior suspension performance can be achieved using the neurocontrol. It is found that the discretized version of the developed training and learning algorithms corresponds to certain well-known existing algorithms.<<ETX>>

Design of observers for two-dimensional systems

This paper extends the familiar 1-D concepts of observer designs to the design of observers for 2-D systems described by Roessers model. Both full-order and reduced-order observers are considered. Extensions of 1-D concepts to 2-D is non-trivial in view of the requirement that state transformations for 2-D systems have to be block diagonal in order to preserve their input-output properties. Another issue addressed in this paper is the required asymptotic stability property of the 2-D observers. To maintain tractability in asymptotic stability analysis, we consider the class of 2-D observers with separable characteristic polynomials. Illustrative examples are provided.

Self-training cognitive preview control for autonomous vehicle path navigation

A description is given of a cognitive preview control strategy for autonomous-vehicle steering and cruise guidance by combining optimal preview control theory with rule-based perceptive cruise command generation. The authors also propose a self-training cognition procedure for determining a suitable perceptive schedule for cognitive cruise and steering control. The control yields humanlike driving action in path navigation. It is an intelligent control that decides the cruising speed, plans its control action, and learns the limitation of its steering control. The strategy is being simulated and tested on an autonomous robotic vehicle testbed which is designed for intelligent control experimentation.<<ETX>>

A Game-Playing Fuzzy Logic Controller for Semi-Active Suspensions

The idea of controlling physical systems with artificially-intelligent control methods has great appeal, and has been investigated heavily in the past. At the very least, they offer the chance to control effectively systems that are highly nonlinear and/or time-varying. This paper describes attempts to fuse game-playing theory and binary-decision-tree pattern recognition methods to build a fuzzy logic parametric controller for a semi-active suspension system. The goal of the controller is to minimize (or defend against) the maximum possible transfer of road interaction forces to the vehicle. Simulation results show that the resulting system can provide the desired control actions to handle smoothly the irregularities in the road profile and the many nonlinearities inherent in the suspension system.

Moving cluster classification technique with lidar traffic monitoring application

A number of methods exist for grouping static data points into clusters. Very few methods, however, consider clustering techniques for grouping dynamic or moving data points. For the purpose of classifying dynamic data into sets of moving clusters, it is necessary to consider the dynamic states, such as position, velocity, acceleration, rotation, etc., of the data. The clustering algorithm must correctly classify the clusters, even if the moving data clusters cross paths and intersect with each other. The paper presents a dynamic clustering method that has been successfully developed and applied to moving laser radar data for an on-board automobile traffic monitoring application. The DCM employs multiple Kalman filters to track the dynamic states of the data points, and a cluster classification and predictor algorithm to identify objects in the information.

Optimal Parametric Control of a Semi-Active Suspension System using Neural Networks

In recent years, there has been a growing interest in controlling both active and semi-active automotive suspension systems with a goal of improving ride comfort and vehicle handling. Many such resulting approaches have used linearized models Of the syspensions dynamics, allowing th use of linear (optimal) control theory. In actuality through, these systems and their optimal control are quite nonlinear. In this paper we propose a novel, yet highly practical alternative to such linearized design methods. This alternate optimal design method consists of a modified A* optimal-path, farward-search algorithm which is highly efficient, together with neural networks. The A* search, using a reasonably accurate system model and a given cost function, establishes te nonlinear optimal parametric control Of the suspension. The neural network, as will be shown, learns this nonlinear optimal control function, and in many ways outperforms the search from which it was taught.

Generalized Optimal Feedback Controllers for Discrete-Time Linear Systems

This paper presents a unified treatment to the problem of optimal feedback control of discrete-time linear constant parameter systems using state feedback, output feedback, dynamical output feedback and observer-controller feedback. A class of generalized optimal feedback controllers is proposed as a basis for obtaining existing results as well as for deriving new results such as the interesting min-max feedback controllers. The essence of the generalized results lies in the minimization of a class of generalized performance measures subject to a flexible controller configuration. The stability aspect of the generalized optimal feedback controlled system is addressed.