Administrator by China Associction for Science and Technology
Sponsored by China Society of Automotive Engineers
Published by AUTO FAN Magazine Co. Ltd.

›› 2019, Vol. 41 ›› Issue (2): 198-205.doi: 10.19562/j.chinasae.qcgc.2019.02.012

Previous Articles     Next Articles

Vehicle State Estimation Based on the Combination of Unscented Kalman Filtering and Genetic Algorithm

Zhou Weiqi1,2, Qi Xiang1, Chen Long1,2, Xu Xing1,2   

  1. 1.School of Automotive and Traffic Engineering, Jiangsu University, Zhenjiang 212013;
    2.Automotive Engineering Research Institute, Jiangsu University, Zhenjiang 212013
  • Received:2017-11-29 Online:2019-02-25 Published:2019-02-25

Abstract: In view of the uncertain situation of the statistical characteristics of process noise and measurement noise in vehicle state estimation, a new adaptive filtering algorithm is put forward by combining UKF algorithm with genetic one for reducing the disturbance of noise to the results of estimation. In order to achieve higher accuracy, a 7 DOF nonlinear vehicle dynamics model is established and by combining ‘magic formula’ tire model, the longitudinal and lateral velocities, tire force and the sideslip angle of mass center are estimated respectively. While UKF algorithm is applied to estimate vehicle states, the genetic algorithm is introduced, and the process noise and measurement noise are optimized based on fitness function to realize the adaptation of noise with the accuracy of estimation greatly enhanced. The results of simulation and road test show that the combination of UKF and genetic algorithms can improve the accuracy of vehicle state estimation with good disturbance resistance

Key words: vehicle, state estimation, UKF algorithm, genetic algorithm, magic formula