Multi-robot path planning in online dynamic obstacle environments based on parallel cooperative strategy optimization algorithm

• Published in: Discover Computing Vol. 28; no. 1; pp. 132 - 24
• Main Authors: Wang, Chia-Hung, Hu, Kun, Wu, Xiaojing
• Format: Journal Article
• Language: English
• Published: Dordrecht Springer Netherlands 01.07.2025; Springer Nature B.V; Springer
• Subjects: Automation; Computer Science; Data Mining and Knowledge Discovery; Data Structures and Information Theory; Dynamic obstacle; Efficiency; Energy consumption; Information Storage and Retrieval; Multi-robot path planning; Multiple robots; Natural Language Processing (NLP); Obstacle avoidance; Optimization; Optimization algorithms; Parallel cooperative strategy; Path planning; Pattern Recognition; Ranking; Robot dynamics; Robotics; Robots; Search algorithms; Static obstacle; Strategy; Task complexity; Unmanned aerial vehicles; Velocity; Parallel cooperative strategy; Static obstacle; Dynamic obstacle; Optimization; Multi-robot path planning
• ISSN: 2948-2992; 1386-4564; 2948-2992; 1573-7659
• DOI: 10.1007/s10791-025-09664-5

The obstacle avoidance path planning for robots has become a critical research focus, especially in the context of addressing complex tasks in dynamic, unstructured environments where unpredictable obstacles and varying conditions present significant challenges. This paper introduces the parallel cooperative strategy reptile search algorithm (PC-RSA) for multi-robot path planning in such dynamic settings. Path planning is crucial for mobile robots, especially in multi-robot systems, requiring solutions that adapt quickly to changes while ensuring efficiency and robustness. Traditional algorithms often struggle with dynamic obstacles and maintaining diversity during search processes. To address these issues, PC-RSA incorporates a parallel cooperative strategy, improving information utilization and balancing exploration with exploitation. The algorithm’s performance was tested using 10-dimensional and 20-dimensional benchmark functions from the CEC2022 test suite and compared with other state-of-the-art algorithms, such as GA, PSO, and RSA. PC-RSA outperformed these methods, ranking first in the Friedman ranking. It was then applied to multi-robot path planning in environments with both static and dynamic obstacles. Simulation results showed notable improvements over the standard RSA, with enhancements of 90.96%, 52.53%, 55.73%, and 62.71% in average path deviation error, average untraveled goal distance, total fitness value, and average execution time, respectively. These findings suggest that PC-RSA could be a promising approach for multi-robot path planning in dynamic environments. Graphical abstract