Vector field based robot navigation using hybrid genetic/simulated annealing algorithm

Abstract

An analytical vector field model of robot workspace was presented. In the model, the vector of resultant field described the most promising direction of robot motion. The model assumed that the edges of every obstacle, which was polygonal, were uniformly charged. It was shown that the resulting repulsive force, which pushing the robot away from the obstacles, could be calculated in closed form. Several factors including the length, the smoothness and the safety of the path require considering in robot navigation. Thus, a hybrid optimization algorithm, HGSA, which incorporated the simulated annealing algorithm (SA) into the genetic algorithm (GA), was proposed to optimize the path through searching the model parameters. The effectiveness of the proposed model was verified by computer simulation in three workspaces with different obstacle distribution. Comparisons between the optimized results show that the hybrid algorithm obtains better path solutions than either GA or SA.