On Line Path Planning for a Circle Mobile Robot Based on Genetic Algorithm

This paper proposes a range sensors information based genetic algorithm(GA) for on line path planning of a circle mobile robot in 2-D unknown environment filled with convex obstacles. The GA uses the current range sensors data and the current robot's position relevent to the goal to decide the robot's next turning angle and moving distance in the world coordination. Some strategies are used for the implementation of on line path planning. One hand the population size equals to the sum of the number of the range sensors and the goal's number and the initial population values are decided by the informations of the sensors and goal too. On the other hand a virtual local environment based on current range sensors data is build in order to further shorten the searching solution time. The previous solution is put into the evolution population pool in genetic operation. The simulation results show the effectivess and feasibility of this method.