High accuracy road vehicle state estimation (using extended Kalman filter

This paper describes the theoretical development and evaluation of the multisensor navigation system for high speed road vehicle. The particular contributions of the theoretical work are in the designing the nonlinear process model that is able to cope with vehicle slip using multisensor data from inertial sensors, odometry, and the D-GPS. The algorithm was evaluated using a vehicle dynamics simulator built to allow the simulation of a wide variety of driving scenarios. The simulation results indicated that the scheme is able to significantly reduce errors in vehicle position and orientation estimates. Besides, the results also show that the scheme allows a slip angle estimation (which is of importance in itself), and accelerometer bias estimation.