Quality-Aware Based Participant Selection For Mobile Crowd Sensing

Mobile Crowd Sensing(MCS),as a new kind of large-scale sensing paradigm,utilizes user-companioned device as basic sensing unit,including smart phones,wearable devices,and so on.These sensing nodes are characterized by wide coverage,real time,mobility and free sharing of data.As a key problem,the measurement and guarantee of data quality play a key role in the MCS application.However,in the process of collecting data,requester often receives unsatisfactory data due to the influence of participants' subjective thought and objective condition.So it's extremely significance that recruiting reliable workers to participate in the sensing task.Based on this,the optimization participant selection mechanism is proposed to assign tasks among the reliable users,as well as data quality measurement and reputation model construction method.The contributions of this paper are summarized as follows:(1)By analyzing large-scale real-life dataset of "KaiTianYan",we define a data quality measurement criteria by combining integrity and accuracy,and thus design its corresponding data quality measurement method and data payment strategy for MCS scenario.(2)To evaluate the reliability of participant,we design a quality-aware reputation model by combining subjective and objective factors,which can be quantified as a direct reflection of data quality participant can provide.We also present a reputation update mechanism by introducing logic regression function.What's more,the state factor and activity factor are used to improve the reputation model.(3)Considering that there are areas where lack of users,we propose different participant selection mechanism satisfying sensing-gain constrained for two scenarios,one is rich at user resources,the other is opponent.In rich resources area,cost and participant scale are set as sensing-gain objectives,and data quality and task coverage are set as a constraint.Then a multistage decision algorithm based on greedy strategy is proposed for solution.In poor resources area,data quality is set as sensing gain target and task is abstracted as 0-1 knapsack problem,a sensing-gain constrained dynamic programming mechanism is proposed.Based on "KaiTianYan" data set,a series of simulation experiments are carried out to verify the performance of the proposed scheme.Simulation results show that the proposed user's reputation model can improve the data quality.The participant selection mechanism proposed for two typical application scenarios can optimize the performance of cost and data quality respectively.