Vehicle state and parameter estimation based on double cubature Kalman filter algorithm

Abstract

Obtaining vehicle status in real-time and accurately during the driving process is of great significance for active safety control of the vehicle. In response to this problem, combining a 7-DOF vehicle dynamic model and the magic formula tire model, the research designed a time-sensitive and robust double cubature Kalman filter (DCKF) observation algorithm. The DCKF algorithm addressed singular value decomposition to optimize the error covariance matrix, and connected driving state observer information of the vehicle to update the observation signal realizing the real-time estimation of the vehicle state. The DCKF algorithm is verified on the simulation platform, and compared and analyzed with the virtual test with CarSim data. The results show that the DCKF algorithm has faster response speed, higher precision of the estimation of the vehicle state, and stronger real-time performance.