Vehicle State Estimation Based on Unscented Kalman Filtering and a Genetic-particle Swarm Algorithm

Abstract

Vehicle state and parameter estimation has gradually become an important way to soft-sense some variables that are difficult to measure directly using general sensors. In the traditional Kalman filtering algorithm, the selection of the process noise covariance matrix and the measurement noise covariance matrix will directly affect the filtering accuracy of the algorithm. In order to improve the filtering accuracy of the filter algorithm to obtain the optimal solution, based on a 7-DOF nonlinear vehicle dynamics model and the Magic formula tire model, a hybrid algorithm containing an unscented Kalman filter (UKF) and a genetic-particle swarm algorithm (genetic-particle swarm UKF) is used to estimate several vehicle key states. Compared with traditional estimator based on UKF and the unscented particle filter (UPF), the simulation and real vehicle test results show that the proposed estimator based on the genetic-particle swarm UKF algorithm has higher accuracy and less computation requirements than the UKF estimator. And also, the proposed hybrid algorithm has superiority on the convergence speed of the optimization than the UPF algorithm. The results of a real-vehicle experiment demonstrate that the proposed hybrid algorithm can be used effectively for solving the vehicle-state estimation problem.