A NOVEL ROBOT PATH PLANNING ALGORITHM BASED ON THE IMPROVED WILD HORSE OPTIMISER WITH HYBRID STRATEGIES

Juntao Zhao, Xiaochuan Luo, and Yong Li

Keywords

Wild horse optimiser (WHO), Sobol sequence, L´evy flight, dynamic self-adaptive factor, opposition-based learning, cubic B-Spline, path planning

Abstract

Metaheuristic algorithms play a pivotal role in addressing the challenges of robot path planning, offering versatile, and efficient solutions. Nevertheless, the standard wild horse optimiser (WHO) has limitations, including limited population diversity during initialisation, constrained global search capability, and challenges in escaping local optima. This paper proposed an improved WHO with hybrid strategies (HI-WHO) to overcome these disadvantages in solving robot path planning problem. The algorithm employs Sobol sequence for uniform population initialisation, integrating the L´evy flight strategy, and dynamic adaptive factor to balance exploration and exploitation. Concurrently, it ensures global search capability and prevents local optima by using the lens imaging opposition- based learning strategy and greedy mechanism. The robustness and effectiveness of the enhanced algorithm were evaluated on a set of 20 benchmark functions. Finally, the improved algorithm, combined with the cubic B-Spline interpolation method, addresses robot path planning in grid map environments, demonstrating its exceptional stability and optimal performance.

Important Links:

Go Back