Dynamic-Model-free vehicle velocity estimation using extended Kalman filter with IMU, steering Angle, and wheel speed sensors

초록

This paper introduces a novel vehicle velocity estimator using an extended Kalman filter (EKF) algorithm and the relationship between velocity and the steered tire (RVST), without relying on dynamic or tire models. Initially, it derives RVST equations under nonslip and slip conditions. These equations are combined to define equality constraints and construct a velocity estimation model called the equations of vehicle motion (EOVM). To achieve accurate velocity estimation without dynamic models, a weighting matrix is designed considering real dynamic characteristics. The EKF prediction model is then built using EOVM and steering angle values, and the measurement model is defined using an inertial measurement unit (IMU) and tire wheel speed sensors. Simulations, conducted under various uncertainties such as tire friction and disturbances, demonstrate the advantages of the proposed dynamic-model-free estimator over existing dynamic-model-based methods.

키워드