A Neural Dynamic Network Pathfinding Algorithm for Autonomous Underwater Vehicles under the Presence of Obstacles
Jiaao Zhao, Song Hee Han, Tao Zhang, Liwen Jia · 2022
A neural dynamic network pathfinding algorithm (NDNPA) is proposed to deal with the path-finding for autonomous underwater vehicles (AUVs). The proposed algorithm is intensified in two aspects: the novel neural activity value network and the improved selection strategy. The impact of the covered positions is introduced into the former. The repetitive path points and energy consumption are reduced simultaneously. In the latter part, in addition to the neural activity values, the navigational directions, the proximity to the target and the extended range of selection are applied to reach the target more efficiently. Finally, compared with the results of other algorithms, the high efficiency of the NDNPA algorithm is proven.