Sensitivity analysis and uncertainty bounding of tire parameters on vehicle handling and stability
摘要
Vehicle handling and stability are significantly influenced by tire characteristics, making the identification of critical tire parameters fundamental to understanding vehicle dynamics. Additionally, the uncertainty associated with these critical parameters poses a major challenge in controller development, as it may lead to vehicle instability under extreme conditions. This paper provides a comprehensive analysis of the impact of tire parameters on handling and stability and proposes an enhanced control strategy considering parameter uncertainties. Initially, the sensitivity analysis of the Magic Formula tire model parameters is conducted. Subsequently, the range of tire parameters and evaluation metrics is further refined using field test data. A co-simulation framework integrating Isight, MATLAB, and CarSim is employed to examine the correlation between tire parameters and vehicle dynamics. Next, a neural network model is utilized to decouple tire parameters, facilitating the quantification of key tire characteristics. Based on this, the general polynomial chaos is applied to model parameter uncertainty, leading to the establishment of a stability margin that accounts for uncertainties. This foundation facilitates the precise definition and implementation of the existing control strategy. Finally, experimental and field test results demonstrate that the proposed method effectively enhances both handling and stability while exhibiting robustness against parameter uncertainties. Compared to the benchmark method, the max-min sideslip angle is reduced by 11.1% and 16.7%.