Abstract:

Crowd sensing is a promising sensing paradigm,which mainly uses a variety of embedded sensors in a large number of mobile devices to accomplish data sensing tasks.One of the key issues in crowd sensing is how to effectively coordinate mobile device users to perform multiple sensory tasks simultaneously.By introducing a new data layer between the sensing task and the user,the similarity of the sensing task and the heterogeneity of the user were effectively utilized,and the joint task selection and user scheduling problems were established on the data layer,aiming at maximizing the social welfare of the whole system.This problem was difficult to solve due to its combinatorial nature and the presence of private information on both the sensing tasks and the users.In order to deal with these problems,a two-sided randomized auction mechanism was proposed,and it was proved that it can satisfy the desirable properties of the computational efficiency,the individual rationality,and the incentive compatibility in expectation.The simulation results show that the proposed stochastic auction mechanism can achieve nearly optimal social welfare,and the social welfare benefits brought by data reuse will increase significantly with the enhancement of task similarity.

Key words: crowd sensing, data reuse, incentive mechanism, auction mechanism