Research On UAV Swarm Cooperative Situation Awareness Consensus Under Granular Computing Perspective

UAV swarm cooperative engagement will be the force to change the “rules of the game” in the future battlefield.Under the complex,highly dynamic and strong antagonistic mission environment,swarm cooperative situation awareness(SA)and situation awareness consensus(SAC)are the basis of swarm autonomous cooperative decision-making and control.However,the studies of SAC at the high level are not sufficient.Meanwhile,traditional multi-UAV cooperation and decision-making methods are difficult to take into account both the adaptability of complex antagonistic environment and the requirement of mission timeliness.Swarm cooperative SAC represents the consensus of UAVs' acquisition and recognition of target situation information in the swarm.The higher the degree of swarm cooperative SAC is,the more conducive to cooperative decision-making.So we can study swarm cooperation methods and information processing mode with SAC.In this thesis,combined with typical scenario of swarm cooperative air-to-ground combat,UAV swarm cooperative SAC is studied based on granular computing,consensus models of group decision making(GDM),etc.The main work and innovations are as follows:(1)Aiming at the problem that the traditional SA three-level model does not fit the characteristics of swarm cooperative SA well and swarm cooperative SAC lacks an analysis model,the UAV swarm cooperative SA model and swarm cooperative SAC three-level model are constructed respectively,and general methods of SAC analysis are designed.The swarm cooperative SAC three-level model includes situation perception consensus(SPC),situation comprehension consensus(SCC)and situation prediction Consensus.(2)Aiming at the problem that the evaluation indices do not fully conform to the requirements of missions and the evaluation methods cannot effectively deal with uncertain information,the evaluation indices of homogeneous swarm cooperative SPC are established,the evaluation method of swarm cooperative SPC based on interval-number processing and the method of swarm cooperative SPC based on three-parameter-interval-number and Heronian operator are proposed respectively.Moreover,evaluation method based on interval-number processing is suitable for the situation of high timeliness and balanced index information;evaluation method based on three-parameter-interval-number and Heronian operator is suitable for the condition of emphasizing the influence of different time,the refinement analysis of index correlation and sufficient time.Experiments show that the evaluation methods can effectively deal with the uncertain situation information and have better performance than the method based on combined weights.(3)Aiming at the influence of swarm network management mode and consensus process in SCC on swarm cooperative SCC formation method,based on node importance assessment method of complex network and the optimal design of feedback mechanism,two kinds of homogeneous swarm cooperative SCC formation methods based on network management mode and GDM consensus theory are proposed.Small-scale homogeneous swarm adopt full-connected peer-to-peer mode,whose time series communication topology is regarded as a time series network with inter-layer similarity,so a swarm SCC formation method based on improved eigenvector centrality and hierarchical feedback regulation is proposed.Experiments show that this method can obtain reasonable evaluation results for network layers with isolated nodes and fully connected network layers,and hierarchical feedback regulation can obtain higher consensus measure and less time overhead.Large-scale homogeneous swarm adopt clustering and hierarchical mode,whose time series communication topology is regarded as a time series network with independent layers,so a SCC formation method based on improved importance contribution matrix and two-stage GDM consensus is proposed.Experiments show that this method has higher accuracy and better performance.(4)Aiming at the difficulty of existing multi-UAV cooperation methods and decision-making methods in considering the adaptability to complex confrontation environment and the requirement of mission timeliness,combined with SAC,the homogeneous swarm cooperation method via SPC,the heterogeneous swarm cooperation method via SPC and the homogeneous swarm cooperation method via SCC are proposed respectively.Correspondingly,homogeneous swarm information processing based on heterogeneous multi-attribute group decision-making(MAGDM)under SPC,heterogeneous swarm information processing based on heterogeneous MAGDM consensus with multiple attribute sets under SPC and homogeneous swarm information processing based on composite heterogeneous GDM consensus under SCC are designed respectively.Combined with the consensus threat assessment of ground targets in heterogeneous reconnaissance UAV swarm cooperative air-to-ground combat,we have the following results.For small-scale homogeneous UAV swarm,a swarm cooperative threat assessment method based on heterogeneous MAGDM consensus is given.Experiments show that this method is dynamic,can effectively process heterogeneous information without information loss and achieve the threat ranking consensus.For large-scale heterogeneous UAV swarm composed of several small-scale homogeneous swarms,a swarm cooperative threat assessment method based on heterogeneous MAGDM consensus with multiple attribute sets is given.Experiments show that this method has better flexibility by transforming heterogeneous consensus into homogeneous consensus and can provide specific preference suggestions for homogeneous swarms.For small-scale homogeneous UAV swarm,when UAV scale,the number of targets and environment complexity increase,swarm cooperative method based on SPC and cooperative information processing mechanism have some problems,such as time-consuming and high communication burden.A threat assessment method for swarm cooperative ground targets based on composite heterogeneous GDM consensus is given.Experiments show that the propose method has more advantages in the cooperative information amount,adjusting information amount and the number of information interactions.(5)Due to the fact that existing distributed cooperation methods of UAV swarm neglect the impact of uncertainty of situation information on swarm cooperation under antagonistic environment and lack quantitative analysis of cooperation performance,analysis indices,such as cooperation time,cooperation information,etc.,are established to analyze the performance of homogeneous swarm cooperative SAC-based methods.The results show that the performance of SAC-based cooperation method is better than that of negotiation-based cooperation method and the performance of SCC-based method is better than that of SPC-based method.