ABS braking on rough terrain

Abstract

This aim of this project may be condensed to the following question:

Is there an improvement achievable in the braking performance of a vehicle on a rough road?

Several follow-up questions arise from the above problem statement:

a) What are the causes of the unsatisfactory stopping time and distance when braking on a rough road and how can they be addressed? b) Can the off-road braking of a vehicle be modelled mathematically? c) What are the criteria used to evaluate the on-road braking performance of a vehicle and can the off-road braking performance of a vehicle be evaluated using the same criteria? d) Can the off-road braking performance be improved without compromising the on-road performance?

An extensive literature survey is done on existing research addressing these four questions. It is found that, although the literature acknowledges that the braking performance of a vehicle deteriorates under off-road conditions, very little has been done to address it. Two main factors influencing the braking performance are identified, namely the ABS algorithm inputs and tyre force generation characteristics. An experimentally validated vehicle model is developed that serves as the basis from which the research question will be addressed. An FTire model is parameterised and used as the tyre model throughout this study. Three measured off-road terrain profiles are used.

The first step in addressing the research question is developing a performance evaluation technique that can easily, quantifiably and visually compare the braking performance of several ABS systems on any road surface, in any condition. The performance evaluation technique considers the stopping distance, longitudinal deceleration, lateral path offset error, and yaw rate error as metrics.

The second step is investigating one of the common assumptions found in ABS algorithms, namely that the roll radius is constant. This is investigated experimentally and it is concluded that the assumption is valid on smooth and rough road surfaces when using the kinematic definition of the roll radius, but invalid when using the kinetic definition of the roll radius.

Investigation of the influence of the tyre force generation characteristics on the braking performance is the third step. It is found that the tyre normal force variation and corresponding suspension force variation correlates closely with the braking performance. A higher suspension force variation is associated with longer stopping distances.

The final step is the development of a three step control strategy that aims to reduce the suspension force variation. This is done by estimating the wheel hop using easy to measure states, predicting the suspension force variation based on these estimates, and finally selecting the ideal suspension configuration.

The control strategy, called the WiSDoM algorithm, was evaluated by doing several simulations on the three off-road road profiles, with different braking points as the only changed variable. The WiSDoM algorithm’s performance was compared with the baseline vehicle performance and found to decrease the stopping distance on all three off-road road profiles, without negatively affecting the stability of the vehicle. The WiSDoM algorithm did not have a significant influence on the braking performance on a smooth road.