Research On Multi-AUVs Hunting Decision-making And Behavior Cooperative Control

In order to effectively enclose the targets found in the deep sea,more and more research institutions are focusing their attention on AUV(Autonomous Underwater Vehicle).The circumvention tasks performed by multiple AUVs can be divided into two stages: decision-making and behavioral coordination of the hunters.These two stages form a complete hunting task in a logical sequence.In the entire mission of the AUV rounding target,the key technical difficulties are: First,what kind of decision-making mechanism to adopt,from the initial overall AUV,to select multiple round-catchers;Second,the selected hunters implement what kind of behavioral coordination strategy that can effectively capture the target in a converging space area..This paper focuses on the decision-making methods of initial overall AUV and the cooperative control strategy of the behavior of the fencers.First,a decision-making method based on improved auction algorithm is proposed.For the traditional auction algorithm,the deficiency of the auctioneer's identity and bidding value lead to the poor security of the selected process and the low efficiency of decision-making.This paper proposes an improved auction algorithm as a decision-making method.Studying the structure and of multi-AUV network communication in the deep sea lays the foundation for decision-making process.The improved auction algorithm,by introducing the principle of geometric center,improved the efficiency of decision-making,enhanced the security and overall rounding efficiency.Second,a behavioral coordination method based on Lions Group algorithm is proposed.Focusing on the problem of behavioral coordination,the continuation of Swarm Intelligence and inspired by the co-hunting behavior of lions in the natural world,a self-organizing lions swarm algorithm was proposed.This article is an in-depth study of the synergy between lion group members and the dynamic game relationship between lions group and prey.A lions group algorithm was proposed and applied to the cooperative control of multiple robots.The behavioral coordination mechanism during the rounding of multiple AUVs was perfected and the clustering intelligence algorithm system was further enriched.Finally,a large number of simulation experiments were conducted.The results show that based on MobileSim and Matlab simulation tools,compared with the traditional auction algorithm,the improved auction algorithm of this paper has higher decision-making efficiency and more stable decision-making process.Compared with Q-learning and SOM,it enhances the robustness of AUV decision-making process.In two-dimensional and three-dimensional environments,the Lions Group algorithm has higher rounding efficiency and better convergence characteristics than fuzzy control and reinforcement learning.