| JOINT ESTIMATION OF VEHICLE STATE AND PARAMETER
BASED ON MAXIMUM CORRENTROPY ADAPTIVE
UNSCENTED KALMAN FILTER |
| Feng Zhang 1, Jingan Feng 1, Dengliang Qi 2, Ya Liu 3, Wenping Shao 1, Jiaao Qi 1, Yuangang Lin 1 |
1Shihezi University
2Northwestern Polytechnical University
3Xi’an University of Posts and Telecommunications University |
|
|
|
|
|
|
|
| ABSTRACT |
|
To address the problem of poor robustness and accuracy of vehicle state and parameter estimation by
conventional Kalman filter in the non-Gaussian environments, a three-degree-of-freedom vehicle model with an improved
Dugoff tire model is established and a joint estimator of vehicle state and parameter is designed using the Maximum
Correntropy (MC) adaptive unscented Kalman filter (AUKF) algorithm in order to simultaneously estimate and identify the
yaw rate, longitudinal vehicle speed, lateral vehicle speed, vehicle mass and rotational inertia. The proposed joint estimator
algorithm was validated by Simulink/CarSim simulation testbed under Double Lane Change and Sine Wave Steering Input
conditions. The results show that MC combined with AUKF (MCAUKF) algorithm has higher estimation accuracy and better
convergence compared to the unscented Kalman filter (UKF) and the MC combined with UKF (MCUKF) in non-Gaussian
environments, and the MCAUKF estimator is more suitable for state estimation and parameter identification of real vehicles. |
| Key Words:
Improved Dugoff tire model, Maximum correntropy, Non-gaussian environment, Vehicle state estimation,
Parameter identification |
|
Corresponding Author.
Jingan Feng , Email. 
Preview 
Full text via DOI 
Download Citation
Print
