Research On Swarm Of Distributed Unmanned Surface Vehicles For Collaborative Search And Targets Positioning

In recent year,marine resources exploration and development,rights protection and law enforcement,and territorial protection tasks in China are increasingly difficult.It is urgent to use marine intelligent equipments to improve China's marine information collection capabilities.The cross platform marine information collection systems consist of Unmanned Surface Vehicles(USVs),Autonomous Underwater Vehicles(AUVs),and Unmanned Aerial Vehicles(UAVs)are becoming the research hotspot.In view of the US Vs swarm subsystem which plays the role of relays in the heterogeneous unmanned system,this paper focuses on the swarm communication,global collaborative search strategy,local collaborative surface targets search strategy and underwater cooperative targets positioning,and carries out the research of distributed USVs swarm collaborative regional search and targets positioning,which provides basic theory and applications for the heterogeneous unmanned systems.Communication is the basis of swarm cooperative control.In order to realize the stability and energy-saving information exchange of dynamic USV swarm communication network,a hierarchical swarm communication network framework is proposed,and the USVlink communication protocol,the networking protocol for constructing the regional communication network and the global communication network are developed.The protocol maintanes and updates the dynamic communication list in the partition,and dynamically constructs the global communication network.In the dynamic partition communication network,a k-medoide clustering algorithm is proposed to optimize communication topology.The algorithm obtains the optimal number of clusters and the best relay node in the swarm within the partition,and takes it as the cluster head to minimize the energy consumption of data transmission between USVs in the swarm system.In the dynamic global communication network,an improved link cost calculation method is proposed.This method comprehensively considers the factors such as the obstruction of the signal,the mobility of the node,the number of adjacent nodes of the cluster head.Ant colony algorithms are used to complete the calculation of the optimal link path of information transmission between partitions,which improves the stability and diffusion of information transmission.On the basis of realizing stable communication among individuals in the swarm system,a multi-objective task model of distributed US Vs swarm collaborative regional search is constructed after a series of search partitions are formed according to the ray method for large-scale multi-reef marine regional search tasks.By introducing the concept of Nash equilibrium,a distributed Nash-BPSO partition allocation algorithm is proposed,which is based on the cooperation among all US Vs in the swarm decision information interaction.This algorithm balances the allocation of regional search tasks in the swarm system.In the case of fully connected swarm communication network,the synchronous parallel iterative optimization strategy in finite time domain is solved,and the optimal cluster global collaborative partition allocation is archieved.After the global level of strategic coordination,USV enters the assigned area to search for the surface targets.Because of the difference of the observed noises from different sensors,a general sensor model of the swarm is constructed by using least squares support vector regression.Simultaneous interpreting the information of different sensors is realized in the form of raster detection confidence,and simultaneous interpreting is made.Grid confidence search graph is generated based on USV local cooperative search.In the process of USVs swarm local collaborative search,aiming at the problems of USVs swarm's cognition of the environment and the high-dimensional dynamic uncertainty and coupling of the results of each USV action,a swarm local collaborative search algorithm based on distributed Markov decision process is proposed by taking the search graph as the collaborative variable.Under this collaborative algorithm strategy,the USVs swarm systems can make the most favorable search decision to reduce the conflict among individuals in the USVs swarm system and minimize the overall uncertainty of information traversing the search area.The USVs swarm system not only needs to search the water surface targets,but also needs to provide certain auxiliary positioning service for AUVs in the heterogeneous unmanned swarm system through underwater acoustic communication.However,underwater acoustic signal is vulnerable to environmental interference and causes data distortion.Therefore,an optimal fusion position estimation algorithm is proposed.In this algorithm,the confidence measure of position estimation is defined to measure twice bit error.By setting the support function,the best fusion confidence distance matrix and the best fusion list are obtained,so as to filter the distorted data.Aiming at the problem of reusing relative distance information between different estimated positions in swarm collaborative positioning,an improved optimal fusion location estimation algorithm based on differential entropy is proposed.The concept of differential entropy is used to calculate the repeatability of each relative distance in location estimation,and the confidence distance is improved.According to the repeatability of time difference information between the location estimation results and other results,the algorithm can calculate the relative distance.The weight value of each position estimation in the final fusion result is calculated to realize the distributed observation fusion of the final position estimation.According to the best fusion list,the transmission frequency of AUVs underwater acoustic positioning signal is improved.