A Genetic- Based Approach to Robot Motion Planning Considering Path Safety

This paper proposes a genetic algorithm for robot path planning by considering both the travelling distance and a safety criterion. Incorporation of the safety issue into path planning is important not only because of the uncertainties in the robot dynamics during path execution but also because of the inaccuracies in the geometric modelling of obstacles. The approach uses a wave front expansion algorithm to build the numeric potential fields for both the goal point and the obstacles by representing the workspace as a grid. The safety value of a node in the grid is defined as the numerical potential for obstacles. A genetic algorithm is developed to search for near optimal paths. Computer simulation results are presented to demonstrate the effectiveness of the algorithm.