| Modern mobile communication devices have developed into powerful and versatile computing platforms.At the same time,a perceptual network constructed by a large number of devices can achieve rapid information acquisition and interaction.Under this background,the Mobile Crowd Sensing network came into being.Mobile Crowd Sensing can take advantage of having a large number of users to realize large-scale data sensing,but how to assign appropriate sensing tasks to mobile users becomes a challenging problem.This article conducts a series of researches on mobile crowd sensing task assignment based on area coverage.The main contents include:This paper proposes an algorithm for determining task observation points based on disk cover(herein referred to as DC-PO).Firstly,based on the historical trajectory data of the users of the sensing platform,the hotspot location in the task area was mined using a typical DBSCAN algorithm,and then the hotspot location was covered using the DC-PO algorithm to complete the task observation point determination.In addition,in order to make the experimental results more vivid,a simulation experimental platform was developed to achieve coverage of the task area with fewer observation points under the conditions of grid distribution and random distribution.In order to push task messages to mobile users in a targeted manner,a recurrent neural network-based location prediction model(herein referred to as LPM-RNN)is used to model the user’s mobile behavior and predict the user’s next location.The simulation experiment uses the Geo Life data set to verify the prediction model,which can predict the user’s position with high accuracy.A task allocation model(herein referred to as MPU-TA)that maximizes the utility of the sensing platform is proposed.It allocates the sensing tasks to mobile users under the constraints of the sensing task time and maximizes the utility of the sensing platform.The formulated optimization problem is converted into a two-stage problem.The first stage uses the FCMPGA algorithm to solve the route planning problem,and the second stage allocates tasks based on the user’s distance to the observation location and the user’s reputation value.Simulation experiment results show that the allocation scheme obtained by MPU-TA is higher lower than the random allocation task without path planning in terms of profit.At the same time,MPU-TA performs task allocation based on the historical perception of users,whichsolves the problem of quality stability issues of perceived data. |
