Research On Fast Networking Methods Of Maritime Multi-agent Under Task Collaboration

With the promotion of carrying out "one belt and one road" policy, development and utilization of maritime resources have been taken of great importance, which is accompanied by the increasingly frequent offshore activities such as marine environment monitoring, search and rescue vessels and marine resources exploration, etc. But the range of human activities and capacity is limited, so an increasing number of agents (such as unmanned aerial vehicles and unmanned boats, etc.) are applied to marine environment, assisting mankind to complete maritime tasks. To ensure the task is accomplished efficiently, studying the construction of a wireless communication network within the agent group is more and more meaningful. By exploring two different types of networking patterns and their representative algorithms, which are network organizing algorithms in Ad hoc network and wireless sensor network, this paper analyzes the differences between maritime and terrestrial networking environment, proposing a type of multi-agent fast network algorithm suited for maritime communication environment, as well as the multi-agent task collaboration distribution strategy. Finally, this paper studies the changes in the motion path when the communication distance between sink nodes changes within the communication area. Specific results are as follows:(1) The application background of this paper is offshore network organizing environment, and agent network is consisted of unmanned aerial vehicles and unmanned boats which are of advanced technique. After analyzing the location of sink nodes, this paper reveals the reason behind maritime network organization different from terrestrial network organization is that the distance between offshore nodes and sink nodes is much larger than land nodes to their sink nodes, which are almost base stations, resulting in rise to a greater information transmission loss during maritime networking process than terrestrial networking process, which affects the communication effects between agents and control center, reducing the quality of task execution. To solve this problem, we adopt multiple UAVs at sea instead of the original fixed base stations as sink nodes of multi-agent fast clustering algorithms, recommending a multi-agent fast clustering algorithm which can be applied in marine environment. The method which cluster head nodes communicating directly with the nearest UAV node is good for reducing the communication distance significantly, thereby reduce the information transmission loss. By comparing with simulation results of other algorithms, it shows that this algorithm has a better performance in maritime networking environment. On this basis, the paper also studies the multi-agent task collaboration distribution strategy. And aiming at the conflicts produced within clusters or inter-cluster during the task allocation process, we put forward the corresponding solutions to solve them.(2) The proposed offshore multi-agent fast clustering algorithm requires UAVs to fly to a specific area as sink nodes, while they should communicate with each other to prevent the communication distance between two of them from being too large beyond communication range and hard to communicate, or from being too small to cause the two UAV nodes collision. Therefore, aiming at the problems of UAVs' collaborative flight, this paper also studies changes in UAV flight paths under three circumstances which are cohesion, alignment and collision avoidance.