Transactions of the Institute of Measurement and Control

This paper demonstrates the use of an extended Kalman filter (KF) as a virtual sensor for non-measurable vehicle states and unknown vehicle parameters. The purpose of obtaining these values is to make them available within the control algorithms of the various automotive stability systems. Based on an extensive four-wheel vehicle model, an estimator is implemented on data from a test vehicle. Using available reference data, the suitability of the extended KF technique as a virtual sensor is demonstrated.

The development of chassis control schemes has been a major area of study for automotive control engineers over the past 30 years. The volume of published literature is large, exceeding 1000 papers. Of this literature, there are 250 examining yaw and sideslip control. Here is a comprehensive review of this field of study to identify the current state of the art and research in yaw rate and sideslip control. The survey shows that there is still a significant research effort needed to address the subjective performance of handling systems, and more research is needed to develop schemes that integrate systems to achieve high-level performance objectives.

The majority of vehicle dynamics control systems currently in production utilize some form of brake or throttle intervention to generate a yaw moment and control wheel slip. Such control systems can be both intrusive and inefficient. The use of active driveline technology is therefore an attractive alternative and recent advances in controlled differential technology have served to make it a potentially viable one. Using simulation results, this paper will demonstrate the power of these devices to influence vehicle dynamics by first proposing a suitable control strategy. This is then used to illustrate how, with perfect actuation, a vehicle's handling characteristics may be modified. The actuator limitations imposed by the two main classes of contemporary controlled differentials are then discussed and imposed on the simulation model. Using the ideal results as a benchmark, the relative merits of each type are then assessed.

Gas struts are widely used in vehicle suspensions. However, they are highly non-linear and conventional linear models are not sufficient to describe their complex behaviour. In this paper, two non-linear gas strut models with different modelling approaches are presented. One is based on grey-box methods using the modular dynamic simulation package Bathfp, where both physical parameters and experimental data are used to model the strut. The other uses black-box methods in Matlab, where artificial neural networks are employed and experimental data are used to train the model. Simulation studies demonstrate that the neural network model is more suitable for real road vehicle simulations, whereas the Bathfp model is more appropriate for detailed system dynamics analysis, especially iso-frequency studies. Since the Bathfp model is able to predict the strut performance when certain parameters in the system are changed and can be easily integrated into different vehicle models, it is a useful design tool for both strut manufacturers and automotive original equipment manufacturers.

Legged robots, as one kind of mobile robots, can be used for tasks too dangerous or difficult for human to perform, eg, planetary exploration, disaster salvation and anti-terrorism action. Consequently, the issues of legged robots, including mechanical structure, stability analysis and control algorithms, have become an important research direction in the field of robotics in recent years. This paper surveys the current status with respect to legged robots, and describes the existent research approaches in terms of mechanical structure, stability analysis and control algorithms of legged robots. To conclude, this paper proposes the problems to be solved and discusses the future development of legged robots.