Worse-case Methodology---Example Applications on Truck Rollover, VSC and Rollover Warning


Active safety technologies (AST) define a group of technologies for vehicle safety enhancement and crash avoidance. AST product development is accelerating rapidly in Europe and Japan, but proceeds at a much slower pace in the U.S. This difference is partly due to business strategy and perceptions of market readiness. A major contributing factor, however, is litigations/liability concerns. To speed up the AST product development, it is necessary to develop an effective evaluation methodology for guaranteed AST system performance under the worst possible scenarios. The worst-case scenario generator algorithm should identify the weak links of the vehicle/control systems and generate worst-case evaluation maneuver in an intelligent manner. A carefully designed test procedure that combines traditional testing and the worst-case testing methods can systematically identify AST performance under common as well as worst-case scenarios, which will accelerate the overall products development cycles of AST techniques.

Given the above motivation, it is clear that a worst-case disturbance generator needs to be developed. The worst-case disturbances will first be generated by linear optimal control and game theories. Depending on the existence, information structre, and goals of the control algorithms, we formulated 4 different linear cases. The linear solutions than serve as initial guess, base on which the final worst-case disturbances are obtained numerically.

Finally, two example vehicle applications will be demonstrated. The first is the rollover and jackknife of articulated vehicles (a one-player example). The second one is the evaluation of an interactive vehicle dynamic system (a two-player example). Application of the worst-case methodology on improved vehicle control and warning systems will be given.