A Multi-Task Algorithm for Autonomous Underwater Vehicles 3D path planning

• Published in: 2020 3rd International Conference on Unmanned Systems (ICUS) pp. 972 - 977
• Main Authors: Hu, Hao, Zhou, Yongjian, Wang, Tonghao, Peng, Xingguang
• Format: Conference Proceeding
• Language: English
• Published: IEEE 27.11.2020
• Subjects: 3D model; Genetic algorithms; MFEA; multi-task optimization; Oceans; Optimization; Path planning; Solid modeling; Task analysis; Three-dimensional displays
• DOI: 10.1109/ICUS50048.2020.9274809

When considering multiple Autonomous Underwater Vehicles (AUV) path planning simultaneously, problems can be seen as multi-task optimization (MTO) problems. The Multifactorial evolutionary algorithm (MFEA) has many obvious advantages in handling multi-task optimization, but it is rarely used in three-dimensional (3D) path planning. Therefore, using the individual gradient based on MFEA (MFEA-IG), this paper solves the global path planning problem of AUV in the 3D environment. We construct a 3D space model, effectively reducing the computational complexity from 3D to two-dimensional (2D). Besides, we also consider the impact of ocean currents on path planning. In this paper, both MFEA-IG and MFEA are used for path planning, and time cost is selected as their optimization target. Experimental results show that the two algorithms can be effectively used for AUV 3D path planning, and MFEA-IG has a better result than MFEA.