A time-optimization algorithm for soccer robot's path-planning in dynamic environment

In order to optimize the robot's motion time, smooth Bezier curves instead of broken line were used to describe the motion path of the robot due to its satisfaction of nonholonomic constraint equation and the ability to control the direction of robot at the target point. A genetic algorithm (GA) was studied to optimize the control points of Bezier curve. The effects (security, length and smoothness of the path) on the motion time were considered in the fitness function of GA. Under the dynamic environment, the potential collision was constructed by predicting the possible position of the obstacles and the avoiding collision robot in every time slice, and the reliability of prediction related to the distance between obstacle and the robot was considered in evaluation of the security factor of the fitness function. The results from the simulation demonstrate the efficiency of the proposed method.