Robust and failure tolerant control of lateral vehicle motion using stochastic robustness

This paper applies stochastic robustness analysis and design (SRAD) to robust, failure tolerant lateral control of a two degree-of-freedom, four-wheel steering (4WS) vehicle. The analysis component of SRAD uses statistical descriptions of uncertainty to determine the probability of not meeting control system design objectives. In the design component, control laws that minimize a stochastic cost function composed of these probabilities are found by genetic search. Specifications for the 4WS vehicle include stability/performance robustness requirements under uncertain load and road conditions, and tolerance to front or rear steering loss. The importance of minimizing performance/robustness trade-offs through careful description of parameter uncertainties and design objectives is stressed, and improving robustness through structured compensation is demonstrated