An accurate fault detection and diagnosis system is of great importance for autonomous vehicles to prevent the potential hazardous situations, and in this paper, a fault detection and diagnosis system is designed, based on hybrid approaches. First, to detect the state faults for the autonomous vehicle, One-Class SVM method is adopted to train the boundary curve which separates the safe domain and unsafe domain. Meanwhile, a Kalman filter observer is designed based on the vehicle kinematic model to predict the current position of the vehicle and get the residuals between prediction and measurement. Interpolation is applied to better infer the residuals distribution, and by checking the normality of the distribution, trajectory deviation in a short period of time can be detected. Further, we design a fuzzy system to distinguish the types of the detected faults based on a modified neutral network, in which a membership function layer is added after the input layer. With the strong self-learning ability of neutral network, the membership function of the fuzzy system is trained with collected data, which greatly reduces the subjectivity of the membership function. Finally, parameters of the membership function are updated by applying black box test techniques. Experiments on the real autonomous vehicle platform validate the effectiveness of these methods and the usability of the fault detection and diagnosis system.

CICTP