Behavior dynamics of collision-avoidance in motion planning of mobile robots

This paper describes the motion planning problems of mobile robots in uncertain dynamic environments with no restrictions on the shape of obstacles. A new approach based on the behavior dynamics of collision-avoidance is proposed to deal with the real-time collision-free motion planning using the local coordinates of the mobile robot. The behavior dynamics of a dynamic collision-avoidance process was modeled, and then the motion-planning problem was transformed into a simple optimization problem with some inequality constraints based on control of the behavior dynamics. The decision-making space of this optimization problem is right the acceleration space of the mobile robot. By solving this optimization problem, the optimal behavior of the mobile robot to avoid obstacles and go to the desired target can be obtained easily. Simulations are given to illustrate the main ideas.