Search Results

The paper describes the application of neural networks to understand the links between test drivers' subjective ratings of vehicle handling and measurable vehicle performance metrics as part of the Foresight Vehicle initiative. The shortcomings of classical linear methods used in previous studies (Crolla et al [1, 2 and 3]) are described along with the processes and developments made in a genetic algorithm based methodology used to find the predominantly non-linear links between subjective and objective handling. The techniques used were designed to make allowances for noise and other distractions inherent within drivers' subjective ratings. Further insights into the preferred ranges of the values of important vehicle handling metrics are presented.

There have been numerous studies of various forms of active suspensions over the past three decades. Most of published literature has reported theoretical studies and outlined the potential advantages in both vehicle ride and handling of such systems over their passive equivalents. One of the systems, which have been shown to have considerable practical potential is a limited bandwidth active scheme based on hydro-pneumatic components. However, in order to exploit the full potential of this arrangement, the control law should include two features; (a) the ability to exploit the wheel-base preview effect in which information at the front suspension of the vehicle is used to improve performance at the rear and (b) the ability to adapt on gain scheduling approach to a variety of different operating conditions. Both features are investigated in the paper using a four degree of freedom model and practical performance benefits are quantified.

The detailed, dynamic properties of dampers are known to influence substantially some of the subtle - and yet nevertheless hugely important - refinement aspects or ride and handling. Despite this, most of the current work on damping characterization relies on steady-state properties and transient aspects are left largely to subjective in-car assessments by test drivers. The paper describes research work aimed at improving our understanding of the transient properties of dampers through mathematical modeling and then attempting to link these properties to detailed aspects of the vehicle ride and handling. Further experimental work is planned to follow later. From a moderately complex mathematical model of a damper, an attempt is made to identify (a) those transient characteristics which are important in influencing the vehicle responses perceived by test drivers, and (b) which design features of the damper control those characteristics.

This paper describes the analysis and design of a hydropneumatic limited bandwidth active suspension system. The design is based on hydropneumatic suspension components taken from an existing European saloon. Mathematical modeling based on the well-known quarter car model is used to compare the performance of the proposed suspension with both the baseline passive system and an idealised fully active system. Practical features, such as actuator time delays and self-levelling controllers, are included in the model of fully active suspension system, and of the limited bandwidth suspension respectively. The controller design for the proposed suspension is based on limited state feedback, and the Nelder-Mead simplex algorithm was used to calculate the gains to optimise the performance index. Overall, the performance of the proposed system appears to be nearly as good as that of the idealised fully active system.

This paper presents the algorithm for a Kalman filter active vehicle suspension design. Based on simulations, two main issues have been investigated, (a) the effects of disturbances from the changes in road input and the variations of vehicle parameters on state observer estimation, (b) the benefits of adaptation of an active suspension to the changes of road input and the variations of vehicle parameters. Simulations showed the significant vehicle performance improvement from adaptation to road input; however, an adaptive Kalman filter is not very necessary.

Torsional motion of a truck driveline system is coupled with other motions of its components. In this paper, a comprehensive model of the truck driveline and body for vibration analysis was developed. Coupling of the torsional vibration of the truck driveline system with the body fore-aft and vertical vibrations was investigated. A mathematical model, including the torsional vibration of the driveline system and the whole body vibrations of the truck, was constructed. The driveline system was modelled as a set of inertia discs linked together by massless springs and the tyre was represented as having massless circumferential band which is elastically connected to the carcass with the bands being subject to longitudinal forces at the road surface. System behaviour at steady and transient runs was developed.

Previous work to examine the performance of a variety of control strategies for a switchable damper suspension system is extended to include an adaptive suspension. The aim of this adaptation algorithm is to maintain optimal performance over the wide range of input conditions typically encountered by a vehicle. The adaptive control loop is based on a gain scheduling approach and two strategies are examined both theoretically and experimentally using a quarter vehicle test rig. For the first strategy, the gains are selected on the basis of root mean square (r.m.s.) wheel acceleration measurements whereas in the second approach the r.m.s. value of suspension working space is used. A composite input is used consisting of sections of a road input disturbance of differing levels of magnitude in order to test the control systems' abilities to identify and adapt efficiently as the severity of the road input changes.