关键词: 萤火虫群优化算法, 路径规划, 非结构环境, 搁浅现象, 移动机器人

Abstract: Against at the problems of oscillation and poor adaptability of robot motion state in the path planning of traditional unstructured environment, a new path planning strategy based on Glowworm Swarm Optimization algorithm of Scene understanding (SGSO) was proposed. The initialization was realized based on the regularity, randomness and generalization of chaotic systems, and golden section method was used for later optimization, which improved the diversity of the population, suppressed the premature and local convergence of the algorithm. And, by introducing scene understanding of glowworm "natural enemy", the selection mechanism of glowworm swarm was optimized to solve the grounding phenomenon of glowworm in the process of tracing under unstructured environment, which enhanced the adaptability and robustness of the algorithm. The simulation results of four test functions show that, the proposed algorithm is superior to the basic Glowworm Swarm Optimization (GSO) algorithm in solving precision and convergence efficiency. The proposed algorithm was applied to the path planning of mobile robots in unstructured environment, the test results show that the planning path based on SGSO was shorter and the corner was more smooth, which could effectively avoid the additional load on power system caused by large angle steering of robot, verifying the feasibility and effectiveness of the proposed algorithm.

Key words: Glowworm Swarm Optimization (GSO) algorithm, path planning, unstructured environment, grounding phenomenon, mobile robot