Vehicle motion state estimation based on WOA-SVR

Obtaining accurate motion state information is the foundation of active movement control systems and advanced driving assistance system. To accurately obtain vehicle motion state information without the high-accuracy dynamic models parameters, this paper proposes a vehicle state estimation algorithm based on Whale Optimization Algorithm-Support Vector Regression (WOA-SVR). A double-layer (longitudinal and lateral) regression estimation structure has been proposed based on analyzing the characteristic of vehicle dynamics. Then, the SVR model is trained on a dataset composed of multiple driving conditions, and the WOA is used to optimize the penalty factor c and kernel function parameter g in the relaxation variable during the training process; Finally, the estimation algorithm was validated through virtual simulation of single line change and frequency sweep tests, as well as ABS braking and double line change actual vehicle tests. The results show that this algorithm effectively improves estimation accuracy and is robust to changes in speed.