Research On Participants Attribute Aware Data Collecting For Mobile Crod Sensing

In Mobile Crowd Sensing(MCS),the ‘Crowd' means a large number of participants,‘Sensing' denotes acquiring the sensing data through their mobile devices,while ‘Mobile' indicates that participants can move randomly without restraint,thus,MCS refers to the wisdom of participants analyzing data,mining knowledge to provide real-time online services.Considering a large number of participants to obtain sensing data in sensing area is the basis for completing sensing task,therefore,designing effective data collection mechanism to accomplish sensing task reliably becomes an urgent problem to be solved.First of all,the concept of MCS is introduced from the cause and data collection process,while analyzing the characteristics of flexibility,efficiency and sociality,the typical application scenarios such as noise monitoring,health care and air pollution monitoring are briefly described.While,summarizes the research hotspots of trust management,incentive,privacy preservation and data collection.Based on this,current data collection mechanisms are classified.Then,a reputation-based collection mechanism is proposed.Specifically,the degree of willingness is measured according to the remaining energy of mobile device and the response time of task.Then,the quality of sensing data is quantified by timeliness,completeness,accuracy and value,and a method for assessing historical reputation of participant's is comprehensively determined.Thus,the server takes historical reputation as an important parameter and updates current reputation by using logistic regression function.In addition,server determines the credibility of sensing data based on current reputation,and effectively selects and receives reliable data while deciding the way of transmission by participant.What's more,participant service ability aware data collecting mechanism is proposed.In particular,participant selects the best sensing task by measuring complexity and desired reward,while server selectively receives sensing data through the minimum total reward and participant's service ability.Therefore,establishing Stackelberg Game model to maximize the utility of participants and server between participants and server,at the same time,updating participants' next service capabilities in conjunction with initial service capabilities and server takes this as an important basisfor whether or not to receive sensing data.Thus,when participants transmit data directly or indirectly to server,server can be guaranteed to perform sensing task efficiently and accurately with the minimum total reward cost.Finally,from the perspective of participant attributes,sum up reputation awareness mechanism and the mechanism with service capability attributes.Based on this,the insufficiency of data quality assessment methods is deeply analyzed,then,looks forward to the future work.