Fuzzy EKF Control for Wheeled Nonholonomic Vehicles

In this paper a new fuzzy extended Kalman robust control system for position and orientation tracking of nonholonomic vehicles with two wheels actuated by two independent DC motors is presented. The problem of robustness and localization are solved simultaneously. About the robustness, some perturbations coming from the outside environment and depending on the contact between the wheels and the ground, involve violations of the nonholonomic constraints. The fuzzy controller of this work is able to obtain a dynamic term of robustness with respect to the perturbations above. However, by using encoders only, the measures of actual position and orientation of the vehicle are with Gaussian noises. Therefore, before the feedback, we use a discrete time extended Kalman filter (EKF), to obtain on-line estimates of the filtered state from the observations of the noised outputs provided from more odometric sensors. Simulations with Matlab 7.0 software confirm the goodness of the proposed control system