Research On The Multiple UAV Collaborative Scheduling For Disaster Data Collection

With the fast development of Internet of Things(Io T),the demand for data collection in Wireless Sensor Network(WSN)is also increasing.Unmanned Aerial Vehicles(UAVs)are widely used in data collection scenarios of wireless sensor networks because of their high agility and flexibility.In particular,in some scenarios with harsh post-disaster environments,UAVs can be quickly deployed for data collected,thereby accelerating rescue speed.However,previous relevant studies have focused on maximizing the data collection effectiveness and maximizing the amount of data collected by UAVs,and insufficient research has been conducted on the scheduling costs and minimizing the completion time of maximum collection tasks for UAVs.In addition,most studies using UAV for data collection ignore the influence of distance on the collection rate and collection time.The constraints of energy consumption and storage capacity in the cooperative scheduling problem of multiple UAVs are also not considered simultaneously.This dissertation first studies the cost optimization problem in UAVs scheduling,using homogeneous UAVs with energy consumption and storage capacity constraints to collect data in all sensing devices,and minimizes the UAV’s scheduling cost by determining the collection location of UAVs,partitioning sensing devices,and fly trajectory of UAVs.Both the collection locations and UAVs fly trajectory have influence on the collection cost,and these two decisions are coupled with each other,thus,it is difficult to solve them directly.This dissertation decomposes the original problem into two sub problems.First,considering the influence of distance on data collection rate,this dissertation proposes the Collection Time Minimization(CTM)problem of UAV,and the Minimum Time Location Determination Algorithm(MTLDA)is designed to solve the CTM problem.Then,based on the collection locations of UAV and sensing device division strategy obtained by CTM problem,considering the constraints of UAV’s energy and storage capacity,this dissertation proposes the UAV Scheduling Cost Minimization(SCM)problem,and a Minimum Scheduling Cost Algorithm(MSCA)with constant approximation ratio is designed to solve the SCM problem and obtain the fly trajectory of UAV.Finally,the simulation results show that,compared with the comparison algorithms,MTLDA reduces the UAV collection time cost at least 37.53% on average,and MTLDA+MSCA reduces the UAV scheduling cost at least 13.21% on average.Secondly,considering the time sensitivity of sensing data in post-disaster rescue scenarios,the freshness of data will largely affect the efficiency of rescue.The higher the freshness of collected data,the more valuable and more helpful it is to assist rescue.Therefore,in order to ensure the freshness of data,this dissertation optimizes the collection location,sensing devices,and collection trajectory of UAVs when the number of UAVs is fixed,in order to minimize the maximum collection task completion time of UAVs.Since this problem is NP hard,this dissertation first considers the impact of UAV collection location on collection time,proposes the Collection Location Deployment Cost Minimization(CLDCM)problem and designs a Minimum Collection Location Determination Algorithm(MCLDA)to minimize the completion time of the maximum collection task with approximate ratio of lnn,where n is the number of sensing devices.Under the condition of fixed UAV collection radius,the proposed algorithm can output the minimum number of UAV collection locations and the optimal sensor division strategy to cover all the sensor devices.Next,given the collection location and sensing device division strategy,the Min-max Collection Task Completion Time(Min-max CTCT)problem is proposed.A Fly Trajectory Planning Algorithm(FTPA)with an approximation ratio of(6+?)is designed,where ? is the search accuracy,to obtain the fly trajectory of the UAV.Finally,the simulation results show that,compared with the comparison algorithms,the proposed MCLDA algorithm can at least reduce the collection location deployment cost by 35.91% on average,and the MCLDA+FTPA algorithm can at least reduce the maximum collection task completion time of the UAV by 17.11% on average.