Abstract
Accurate information on tire road friction coefficient (TRFC) is essential to autonomous driving systems. In this paper, a fault-tolerant estimation scheme is proposed to estimate TRFC in the case of missing measurements. First, a fault-tolerant unscented Kalman filter (FTUKF) is developed for estimating longitudinal and lateral tire forces in the condition of sensor signal loss. Then, longitudinal and lateral TRFCs are estimated separately with FTUKF based on tire forces information. Next, an event-driven multi-model fusion method based on the degree of data loss is designed to perform a weighted fusion of longitudinal and lateral TRFCs to further improve the estimation accuracy. Experiments with different working conditions are performed to demonstrate the validity of the fault-tolerant estimation framework. The results illustrate that the designed approach has higher estimation accuracy and strong adaptability under various roads.
Original language | English |
---|---|
Pages (from-to) | 1066-1078 |
Number of pages | 13 |
Journal | IEEE Transactions on Intelligent Vehicles |
Volume | 9 |
Issue number | 1 |
DOIs | |
Publication status | Published - Jan 2024 |
Keywords
- Autonomous driving systems
- event-driven fusion method
- fault-tolerant unscented Kalman filter
- tire road friction coefficient
ASJC Scopus subject areas
- Automotive Engineering
- Control and Optimization
- Artificial Intelligence