Research On Key Technology Of Task Assignment Based On Lyapunov Optimization In UAV CrowdSensing

With the rapid development of information technology and the increasing popularity of mobile smart devices,Mobile CrowdSensing(MCS)is gaining attention as a new mode of perception and processing.MCS uses the smart devices that users carry around to form interactive and participatory sensing networks to distribute sensing tasks and obtain sensing information to solve large-scale complex sensing and computation problems.Compared to previous sensing approaches that relied on specialist technicians and special sensing devices,MCS has many advantages,such as wide coverage,flexible deployment,multi-source heterogeneity and scalability,and is therefore widely used in many fields such as environmental monitoring,traffic monitoring and indoor positioning.As sensing tasks become increasingly complex and diverse,traditional mobile crowdsensing scenarios become more difficult and dangerous,and sensing units made up of ordinary users gradually fail to meet the demands of the task.So Unmanned Aerial Vehicles(UAVs),’special users’,are introduced to participate in MCS systems.To ensure the long-term stable operation of the UAVs crowdsensing system,the platform needs to consider the stability of the UAV’s cache space when making task assignments(i.e.UAV’s data collection rate should not exceed its data upload rate).In addition,the accuracy of the sensing tasks(i.e.the coverage rate of each task)is an important metric for evaluating the quality of task completion in MCS system,so the platform needs to guarantee the completion amount of each task when assigning tasks.In the overall sensing process,energy consumption is a factor that cannot be ignored when the UAV performs sensing tasks,so it is necessary for the platform to perform path planning for the UAV to reduce system sensing cost.In this context,how to ensure the stability and accuracy while performing efficient task assignment and path planning to reduce the sensing cost has become a key issue.To solve the above problem,this paper designs a Lyapunov optimization-based UAV task assignment algorithm,as follows.In this paper,the stability constraint of UAVs and the accuracy constraint of tasks are modelled as the UAV data processing queue and task virtual accuracy queue respectively,and the dual constraint is satisfied by maintaining the stability of the queues.The paper then extends the traditional single-queue framework based on the original Lyapunov optimisation model by forming a joint queue(i.e.a double queue)with the data processing queue and virtual accuracy queue,and designs a UAV task assignment algorithm for task assignment and UAV path planning,which finds a tradeoff between minimising sensing cost and ensuring queue stability.In addition,the performance of the algorithm is analysed at the theoretical level,and an upper bound on the sensing cost of the system and the queueing delay backlog is proved by rigorous mathematical reasoning.Finally,this paper evaluates the performance of the algorithm by implementing extensive simulations on three real-world motion trajectory datasets,Roma/taxi,Epfl/mobility and Feeder,which are commonly used in the real world.Based on the experimental results,it is clear that the proposed algorithm achieves a better compromise in terms of UAV stability,task accuracy and sensing cost compared to the baseline approach.