AUV path planning in a three-dimensional marine environment based on a novel multiple swarm co-evolutionary algorithm

Zhilei Liu,Dayong Ning,Jiaoyi Hou,Fengrui Zhang,Gangda Liang

DOI: https://doi.org/10.1016/j.asoc.2024.111933

2024-07-14

Applied Soft Computing Journal

Abstract:Autonomous underwater vehicles (AUVs) are rapidly advancing in ocean exploration. High-performance path planning techniques are essential for AUVs. Path planning for AUVs is a multi-faceted challenge, necessitating careful consideration of safety, energy consumption and the influence of sea currents to ensure the development of a high-quality trajectory that satisfies all operational criteria. Existing path planning algorithms have problems such as incomplete consideration of influencing factors, computational complexity and weak applicability. Therefore, we propose a swarm intelligence optimisation algorithm based on multiple swarm co-evolution (MCO) to address the issue of AUV path planning in a three-dimensional marine environment. First, a three-dimensional marine environment model and the corresponding path evaluation mechanism are established. Second, the MCO rule is established to ensure that the MCO has balanced exploration and exploitation capabilities. A shared dynamic optimal particle between populations is introduced to ensure information exchange between populations. In addition, the cross-integration mutation strategy has been proposed for promoting the fusion of two populations of superior genes to ensure the inheritance of superior paternal genes to the offspring. Finally, four comparison experiments are designed, and the experiments compare eight commonly used intelligent search algorithms and improved versions. The results of the experiments proved that the MCO has excellent three-dimensional marine environment path planning capability, with robustness and search capabilities superior to other swarm intelligence optimisation algorithms.