An Improved Dynamic Arc Fitting Trajectory Tracking Method

To improve the speed and accuracy of path tracking of a skid-steered robot, an improved algorithm based on the dynamic arc fitting method is presented. A dynamic look-ahead distance model is formulated to cut the global path when the points at which the heading angle change drastically are detected. In addition, the obstacle potential field model with the trajectory collision detection equation is considered to ensure the safety of the fitted trajectory. The average arc radius model, radius uniform variation model and hard constraints of attitude adjustment range are formed to improve the moving speed and stability of the robot. Furthermore, a prediction and compensation model of slip error is modeled to reduce the trajectory tracking error. Finally, the validity of the proposed methods is verified by simulations and experiments.