Research on trouble diagnosis method of braking anti-lock system based on vehicle data flow

Abstract

Trouble diagnosis method of anti-lock braking system based on vehicle data flow was proposed. The selection principles of trouble diagnosis parameters and trouble modes were analyzed firstly, then 3 trouble modes and 9 parameters including four wheels’ speed, vehicle speed, brake pedal position, steering wheel angle, lateral acceleration and yaw rate were selected based on SAE J1939, 48 sets of test data were obtained by vehicle trouble simulation test. The test data were pretreated to form 32 sets of training samples and 16 sets of test samples, and then neural network models were trained and validated. The accuracy of the test results reach 100% by adjusting and optimizing smoothing factor values. The results show that the neural network model based on vehicle data flow can achieve accurate trouble diagnosis.