A Hybrid Modified ABC-PSO Algorithm for Optimal Robotic Path Planner

Abstract

Path planning is one of the most fundamental problems that must be dealt with before the mobile robot can navigate and explore autonomously in any environment. A good path-planning algorithm can save time and reduce mobile robot wear and capital investment. Path computing time and average path length are important factors over cost functions that reflect the algorithm's effectiveness, such as power consumption or average trip time. The Artificial Bee Colony (ABC) represents one of the most important global search algorithms. The main problem with ABC is that it suffers from a slow convergence rate due to lousy exploitation and tends to get trapped in the local minima. This paper proposes and evaluates a new robot path-planning algorithm named Modified Artificial Bee Colony (MABC). MABC algorithm design is based on modifying the ABC algorithm by cross-layer design between ABC and Particle Swarm Optimization (PSO) algorithms. The MABC is different from the original ABC algorithm in that it modifies the original one to use PSO's exploitation rather than its exploitation. On the other hand, the PSO algorithm has better exploitation but poor exploration characteristics. The evaluation and analysis were performed for several performance metrics and under different evaluation scenarios. It has been observed from the results that the MABC-PSO outperforms the original ABC with respect to average path length and convergence time which leads to improving the planning of the path.