Robust Fault Classification and Handling for Longitudinal Vehicle Control

Abstract

In this paper, a method of analyzing the performance loss caused by faults that effect the longitudinal control of an automated vehicle is presented. First, a hybrid longitudinal controller based on Dynamic Surface Control (DSC) is developed to provide robust stabilization. Then, ellipsoidal approximations of the tracking error bounds are obtained via convex optimization for the nonlinear closed-loop system. The bounds are then used to design a fault classification scheme and fault management system based on the amount of performance loss as measured by a quadratic Lyapunov function. Simulation results of the fault tolerant controller are shown for both single and multiple multiplicative faults.