Research On Autonomous Cooperative Search Using UAV Swarms

One of the main applications of unmanned aerial vehicles(UAVs)is to carry out coverage reconnaissance and target search in unknown areas.Compared with single UAV,UAV swarms has more advantages in this aspect.In consideration of complex situations such as the actual task type,environmental factors,and the dynamics of the UAV,the realization of collaborative control among UAVs to complete reconnaissance and target search tasks with high efficiency and low cost is the main problem that UAV swarms need to solve,which is also hot and difficult issue in this field in recent years.The main works in this paper are summarized as follows.In this paper,reconnaissance and target search tasks in the unknown region are decomposed into coverage problem and search problem.An area coverage model of multi-UAV based on central Voronoi configuration is established,which realizes local optimal deployment and solves the fundamental problem of multi-UAV task coordination.The characteristics of the model are as follows:Firstly,UAV swarms adopt completely distributed control,and its dynamic system is continuously evolved based on the(sub)optimal solution.Secondly,in order to conveniently select different sensor function types according to the actual application scenario,distribution density function and different sensor performance functions of time-space changes are defined,so that the model can be used more flexible.Finally,the model can deal with the non-convex environment and heterogeneous problems,that is,the sensors in the model could be in different forms of carriers(such as UAV,unmanned vehicles,etc.)to ensure good compatibility.A multi-UAV cooperative search strategy based on probabilistic map information updating and fusion is designed.For a group of UAVs with limited sensing and communication range,the search for multiple ground targets in an unknown environment is realized.Firstly,each UAV communicates with neighboring group members within the communication range and shares local target information,and controls its own behavior in a distributed manner.Secondly,the influence of time-varying detection caused by altitude change of the UAV on the uncertainty information of the probability map in the three-dimensional dynamic environment is analyzed,and the information variation factors between the area units caused by environmental changes such as target movement and random disappearance/appearance are introduced to improve the search model.Thirdly,an optimization function is designed in the search algorithm,which combines task efficiency and flying cost to improve regional coverage and target detection performance.Under the assumption that each UAV is located at any position in the unknown area,the traditional carpet or spiral search cannot be adopted.Therefore,compared with the conventional path planning method with prior information and a certain destination,the method in this paper is more difficult and more universal.Finally,according to the above theoretical work,based on Matlab simulation platform,we develop multi-UAV cooperative target search core algorithm,performance comparison function etc.using C language.This paper simulates and verifies 2 sorties,6 sorties and 18 sorties of UAV perform reconnaissance and search tasks in the unknown environment.In addition,a comparative experiment of search ability of UAV swarms is carried out in unknown environment with and without obstacles,which verified the robustness of the system to the complex environment.