Multi-AUV Cooperative Hunting Control with Improved Glasius Bio-inspired Neural Network

Abstract

Cooperative hunting with multiple Autonomous Underwater Vehicles (AUVs) not only needs the AUVs to cooperate, but also demands real-time path planning to catch up with evading targets. In this paper a time-based alliance mechanism to form efficient dynamic hunting alliances is proposed. After that, during the active hunting stage, an improved neural network model based on a Glasius Bio-inspired Neural Network (GBNN) is presented for path planning to immediately achieve tracking of an intelligent target. This study shows that the improved GBNN model has good performance in real-time hunting path planning. From the simulation studies as described in this paper, both the hunting alliance formation mechanism and the proposed real-time hunting path planning strategy show their advantages. The results show that the improved GBNN model proposed in this paper can work well in the control of multiple AUVs to hunt for intelligent evading targets in environments containing obstacles.