In order to solve the frequent occurrence of road vehicle functional safety problems and the imperfect fault diagnosis mechanism of electric vehicle controller, this paper proposes a vehicle control unit (VCU) functional safety fault diagnosis strategy based on Random Forest and Hidden Markov Model. The vehicle operation process is described as a Hidden Markov stochastic process, and feature screening is carried out with the help of Random Forest. The implicit relationship model between the functional safety fault of the VCU and the change of the characteristic signal under the vehicle fault state is established and the fault diagnosis is solved. Using the constructed HIL simulation test environment, the diagnosis and verification of six functional safety faults of accelerator pedal to ground short circuit, accelerator pedal opening signal is not accurate, brake pedal to ground short circuit, brake pedal opening signal is not accurate, application layer software logic fault and MCU power supply abnormal are realized. The results show that the proposed diagnostic strategy has an average diagnostic rate of 90.1%, which can effectively diagnose different types of functional safety faults.
Functional Safety Fault Diagnosis Strategy for Electric Vehicle VCU Based on Random Forest and Hidden Markov Model
2023-12-14
1282882 byte
Conference paper
Electronic Resource
English
Fault Diagnosis Method Based on Hidden Markov Model
Springer Verlag | 2016
|Diagnosis of mechanical fault signals using continuous hidden Markov model
Online Contents | 2004
|British Library Online Contents | 2017
|