Adaptive estimations of tyre–road friction coefficient and body’s sideslip angle based on strong tracking and interactive multiple model theories

Vehicle dynamic states and parameters, such as the tyre–road friction coefficient and body’s sideslip angle especially, are crucial for vehicle dynamics control with close-loop feedback laws. Autonomous vehicles also have strict demands on real-time knowledge of those information to make reliable decisions. With consideration of the cost saving, some estimation methods employing high-resolution vision and position devices are not for the production vehicles. Meanwhile, the bad adaptability of traditional Kalman filters to variable system structure restricts their practical applications. This paper introduces a cost-efficient estimation scheme using on-board sensors. Improved Strong Tracking Unscented Kalman filter is constructed to estimate the friction coefficient with fast convergence rate on time-variant road surfaces. On the basis of previous step, an estimator based on interactive multiple model is built to tolerant biased noise covariance matrices and observe body’s sideslip angle. After the vehicle modelling errors are considered, a Self-Correction Data Fusion algorithm is developed to integrate results of the estimator and direct integral method with error correction theory. Some simulations and experiments are also implemented, and their results verify the high accuracy and good robustness of the cooperative estimation scheme.