Research On Distributed Collaborative Decision Making Method For Unmanned Surface Vehicle Group For Escorting Task

The International Maritime Bureau’s "Piracy and Armed Robbery at Sea Report(2020)" shows that there have been up to 200 pirate attacks in recent years.Using multiple unmanned surface vehicles to perform escort tasks instead of traditional ships can save escort costs and avoid casualties,which is a powerful way to maintain marine safety.Collaborative decisionmaking technology improves the system’s working ability through orderly division of labor,and is the key technology of multiple unmanned surface vehicles collaborative system.This paper takes the escort mission of large merchant ships as the research background,and the main contents are as follows:Firstly,the common multi-agent cooperative structure is analyzed,and the multiple unmanned surface vehicles distributed cooperative structure for escort tasks is designed,including subsystems such as situation processing,cooperative decision-making,and task execution.The escort process of large commercial ships is sorted out,the situation processing rules are determined,the threat of moving targets on the sea surface is determined qualitatively and quantitatively,and the objective function of intercepting specific ships when unmanned boats are escorting is constructed.Secondly,the clustering problem when the number of maritime threat targets is large is studied,a clustering scheme based on the location of interception points is proposed,and the K-means algorithm is improved to improve the adaptability to outlier interception points.The particle swarm algorithm is used to realize the path planning of the unmanned surface vehicle to the interception point,and the initialization method of the particles is designed to improve the interception effect of the threat target.The superiority of the improved K-means threat target clustering method and the particle swarm-based interception path planning method is verified by simulation.Then,aiming at the collaborative decision-making problem of multiple unmanned surface vehicles in escort missions,the unmanned surface vehicles and tasks are marketed,and a collaborative mechanism based on merchant pricing algorithms is established to realize collaborative decision-making among boats.Aiming at the problems of too many negotiations and low efficiency of the traditional merchant pricing algorithm,the price reduction stage is introduced to improve the merchant pricing algorithm.A discount strategy for neglected mission to improve its attractiveness to unmanned surface vehicle buyers is designed,so as to achieve the purpose of conflict resolution.The superiority of the multiple unmanned surface vehicles collaborative decision-making method based on improved merchant pricing is verified by simulation and results analysis.Finally,the collaborative decision-making problem of multiple unmanned surface vehicles with distributed network structure and non-fully connected communication topology is studied,the central role of merchant is cancelled,the consistency mechanism of merchant pricing algorithm is analyzed,and the communication protocol and consistency rule table are designed to make the bidding of unmanned surface vehicle converge every round.Simulation results show that the consistent merchant pricing algorithm can achieve the same distribution results as the centralized algorithm,and the consistency mechanism is effective.