Task Allocation Mechanism Of Data Collection In Mobile Crowd Sensing Networks

Mobile Crowd Sensing(MCS)is a novel emerging paradigm that leverages the recent surge of sensor-equipped smart mobile terminals(e.g.smartphones,tablets and intelligent wearable devices)to collect information.Users who carried smart mobile terminals can form MCS networks by collaborating with each other.Then they can complete large-scale social sensing tasks together.Compared with wireless sensor networks,MCS has advantages of lower data collection costs,easier system maintenance,better scalability and other.These features make MCS develop rapidly.However,lots of problems have shown when MCS is applied to complex sensing scenes.The most important problem is that MCS system needs to process high-dimension and large-scale data.For large-scaled data collection,sensing users are not evenly distributed;some areas are crowd with many users far out weight the needs.If the sensing states of users are not controlled,redundancy users,which performing unnecessary sensing tasks,would cause a serious energy consumption.Also,the limitation of data types of sensing users would result in a failure of supporting high-dimension data collection tasks.While a large deployment of sensors can be a huge cost.In order to ensure the coverage requirements of applications and reduce the energy consumption during data collection,and to reduce the costs and complexity of high-dimension data collection,this thesis studied the energy consumption control mechanism of sensing and the task allocation mechanism for high dimensional data collection.The main contents of this thesis are listed as follows.1.Designed the energy consumption control mechanism of sensing.This mechanism controls the sampling behavior of the mobile terminal based on the user's movement rule and the sampling requirement of the region.If the number of active users of one grid in the area exceeds the number of needed active users,the central server would inform some users in that grid to close their sensors to avoid unnecessary energy consumption.2.High-dimension task allocation mechanism.The exists task allocation methods are applied to the tasks which are relatively simple specific,or terminals who default have comprehensive sensing ability.However,the terminals cannot meet the requirement of complex sensing task.There need to reduce the dimension of high-dimension data,then distribute the sub-tasks to the users to complete the sensing task.3.A trade-off algorithm for the cost of data collection and the fairness of user participation is proposed.The algorithm is based on the user's information attributes,it balances the selection probability of users with considering the sensing ability and the quotes of them,then selecting an optimal user sets for sub-tasks.Eventually,the algorithm can reduce the cost of data collection while improving fairness of user participation.This thesis proposed corresponding models and methods from two aspects,energy control of sensing and high-dimension task allocation.The effectiveness of them have been demonstrated by theoretical analysis and simulation analysis.The study in this thesis provides a theoretical basis for data collection and task allocation of complex MCS application,which has a certain reference value.