Hibrid Adaptive/EKF Motion Control and Data Fusion for Ground Vehicles with Kinematical and Dynamical Uncertainties

This paper considers the motion control problem of ground vehicles with nonholonomic constraints and parametric uncertainties both in the kinematic and in the dynamic model. The presence of uncertainties above is treated using adaptation laws where the Lyapunov's stability of the position and orientation errors is proved. Now, if the feedback signals are position and orientation provided by incremental encoders only, then noises of the odometric sensors above can damage the control in terms of difference between the desired and the actual motion of the vehicle and in terms of performances of the parametric adaptation. So an extended Kalman's filter (EKF) is inserted in the feedback for measuring and reducing the odometric noises above. Based on fusion of data provided by multiple proprioceptive sensors (i.e. incremental encoders, vector compass and position sensor), the EKF estimates the state of the vehicle, i.e. position and orientation, to obtain a filtered feedback signal. The adaptive control and the on-line EKF lead to the hybrid adaptive/EKF control of this paper. The control algorithm efficiency is confirmed through simulation experiments.