Improved Artificial Fish Swarm Algorithm Approach to Robot Path Planning Problems

Abstract

Path planning is one of the research hotspots in the field of robotics and it has certain difficulties. A novel intelligent optimization algorithm, artificial fish swarm algorithm (AFSA) is employed to solve this problem. When establishing the environmental model by polyhedral terrain technique, we present a method to determine the number of fence lines. Especially to overcome the shortcomings of original AFSA, such as low convergence rate and poor ability of balancing exploitation and exploration, an improved artificial fish swarm algorithm (IAFSA) is proposed. The basic idea of proposed algorithm is to divide the whole population into two sub groups, and different adaptive strategies are applied to each of them respectively. These two subpopulations evolve independently and individual migrations are conducted regularly to increase the population diversity and improve convergence rate of the algorithm. The decision probability is also introduced to reduce its computational complexity. In addition, differential evolution is hybridized with AFSA in order to make proposed algorithm escape from local optima. Then IAFSA is employed to solve robot path planning problems. Compared with the original AFSA, results show that proposed IAFSA can improve overall optimization performance. And the feasibility and effectiveness of IAFSA are verified.