Research Of Prediction And Incentive Algorithms In Human Intelligence Sensing System Under The Concept Of Participatory Sensing

Participatory Sensing,which is as a subsidiary of pervasive computing,has been developing faster with the increasing number of smart mobile devices than before.If the concept of "All Thing Sensors" is applied on the Participatory Sensing,the human also can be treated as a kind of sensor which can generate sensing data,upload or share them with others.That is the concept of Human Intelligence Sensing which the paper proposes.In the Location Related Expectation Maximization prediction algorithm,the basic idea of Expectation Maximization is used.The Location Related Binary Sensing Data is built,and the problem is formulated as solving the maximum likelihood function which has several latent variables.The location relationship is considered at the same time,and the extensive simulations and a real world case proves the effectiveness of the algorithm.In the Imperfect Public Information Repeated Gaming Incentive Algorithm,the Green-Porter Model is used to define the noncooperation in the Participatory Sensing System.The Price-Trigger Nash equilibrium is proved based on the model,and the statistic Nash equilibrium,which is benefit for improving the performance of the network,among all participators is also proved.The incentive algorithm is designed according to all these models.The extensive simulations prove that the Imperfect Public Information Repeated Gaming Incentive Algorithm can increase the delivery ratio,decrease the average delay,and accelerate participators to give the lowest price they can accept.The paper mainly focuses on the prediction and incentive algorithm under the Participatory Sensing environment,and it plays a positive effect on studying the Participatory Sensing and Human Intelligence Sensing.