A Motif-based Mission Planning Method For UAV Swarms

The emergence of UAV(unmanned aerial vechile)has brought about great changes in the way of operation.Compared with traditional manned aerial vehicles,UAVs have the advantages of zero casualties,long continuous combat time and low cost of full life cycle.Therefore,UAVs are playing an increasingly important role in warfare.The mode of operation begins to change from traditional manned warfare to unmanned warfare.Faced with highly confrontational,highly uncertain and highly dynamic battlefield environment,the operational mode of UAV has gradually developed from single-platform operation to cluster networked cooperative operation.Mission planning of UAVs cluster has become a hot and difficult issue.At present,experts at home and abroad have studied the mission planning of UAVs from different perspectives,and have achieved fruitful results.However,these mission planning methods do not take into account the actual communication constraints between UAVs.Due to the restriction of communication technology of UAVs,such as bandwidth,frequency and so on,as well as the influence of complex combat environment,the communication of UAV cluster is greatly limited.Therefore,UAV cluster needs dynamic allocation with the change of mission.How to do mission planning for UAV swarms under limited communication has become a hot topic and a difficult problem.In view of this situation,motifs are introduced as basic units of UAV swarm configuration.A motif is an infrastructure unit that only requires few connections.On the basis of module-based collocation,this work presents a mission planning method for UAV swarms based on motif reconfiguration.We introduce a general framework planning model with three layers of structure.Firstly,a mission is decomposed into a series of logical tasks by the goal decomposition method.Then,these subtasks are mapped to the type and number of motifs required for these tasks through the capability requirements and information requirements.With the multi-dimensional dynamic list scheduling method,the priority execution sequence of tasks and the types and numbers of motifs needed to complete each task are taken as input,and the mission planning scheme is automatically generated.The priority execution sequence of each task corresponds to a mission planning plan.The explosion of the combination of task sequences causes the exponential increase of computational complexity.It is a huge computational complexity problem to find the optimal planning scheme which meets the requirement from a large number of feasible mission planning schemes.In this paper,three indicators are used to evaluate the quality of mission planning: mission completion time,communication connection change and the number of UAVs used each time.The problem of selecting qualified task planning schemes from a large number of feasible schemes is actually a multi-objective optimization problem.The optimization variables are the priority execution sequence of tasks,and the optimization objectives are three indicators for evaluating the schemes.This multi-objective optimization problem has a large number of feasible solutions,complex solving process and time-consuming,which does not meet the requirements of battlefield and timely mission planning.In order to improve the computational efficiency and reduce the computational time,this work applies NSGA-III to solve this problem.Finally,the feasibility and validity of the proposed motif-based mission planning method are preliminarily verified by a case of UAV cluster capturing airport and port.