Research On Task Assignment Mechanism In Mobile Crowd Sensing

Mobile crowd sensing uses smart mobile devices carried by mobile users as sensing nodes to collect all kinds of data and complete large-scale and complex sensing tasks.It has the advantages of wide coverage and low deployment cost,and has broad application prospects in intelligent transportation,environmental monitoring,smart healthcare and other scenarios.Task allocation is an important part of mobile crowd sensing.The main purpose is to assign sensing tasks to appropriate participants,so as to improve the quality of task completion.Due to the variability of sensing scenarios,complex task requirements,and incomplete consideration of the attributes of tasks and participants,the current task allocation methods are not applicable.On the other hand,for large-scale traffic flow monitoring scenarios,the division of task importance and the evaluation of participant quality also need to be solved urgently.In order to solve the above problems,this thesis studies the task allocation mechanism of mobile crowd sensing,and proposes two mobile crowd sensing task allocation methods.The specific research content is as follows:(1)Task assignment method based on spatial-temporal information and participants’ positivity.The current task allocation methods rarely consider the sensing duration of the task,and are not suitable for the allocation scenario where the idle time of the participants is limited,and it is difficult to allocate long-term sensing tasks.In this thesis,considering the sensing time,location entropy and deadline of the task,a task evaluation mechanism based on spatial-temporal information is proposed.According to the probability of participants accepting tasks and historical task completion,a participant positivity measurement model is constructed.Combining the task evaluation mechanism with the participant positivity measurement model,a task allocation method based on spatial-temporal information and participant positivity was proposed.Aiming at maximizing the number of tasks allocated,the task allocation problem is transformed into a maximum weight binary matching problem,which is solved by Kuhn-Munkres algorithm.(2)Task assignment method based on dynamic programming for traffic flow monitoring.For large-scale traffic flow monitoring tasks,the importance of monitoring subtasks is different in different regions and different periods of time,which increases the difficulty of task allocation.According to the time-varying characteristics and regional characteristics of traffic flow monitoring tasks,this thesis constructs a sub-task importance partition model to evaluate the impact of each sub-task on the overall task completion quality.In terms of participant selection,a participant quality evaluation model was constructed according to the quality of data submitted by participants and the coverage rate of monitoring duration.Combining the subtask importance partition model with the participant quality evaluation model,a dynamic programming based task allocation method for traffic flow monitoring is proposed.Taking the overall task completion quality as the optimization objective,the task allocation problem is transformed into a knapsack problem,which is solved by dynamic programming method.The simulation results show that the task assignment method based on spatialtemporal information and participant positivity has better performance in the number of task allocation,participant time utilization and participants’ positivity.The task assignment method of traffic flow monitoring based on dynamic programming is superior to similar algorithms in terms of overall task completion quality,average participants’ quality and budget utilization.