Incentive Mechanisms for Crowdsensing

Crowdsensing is a popular method that leverages a crowd of sensor users to collect data. For many crowdsensing applications, the collected raw data need to be preprocessed before further analysis, and the preprocessing work is mainly done by the crowdsourcer. However, as the amount of collected data increases, this type of preprocessing approach has many disadvantages. In this article, we construct monetary-based incentive mechanisms to motivate users to preprocess the collected raw data for the crowdsourcer. For two common crowdsensing scenarios, we propose two system models, which are the single-task-multiple-participants (STMP) model and the multiple-tasks-multiple-participants (MTMP) model. In the STMP model, we design an incentive mechanism based on game theory and prove that there is a Nash equilibrium. In the MTMP model, we develop an incentive mechanism based on an auction and demonstrate that the incentive mechanism has the desirable properties of truthfulness, individual rationality, profitability, and computational efficiency. Furthermore, the utility maximization problems of the crowdsourcer and users are simultaneously considered in our incentive mechanisms. Through theoretical analysis and extensive experiments, we evaluate the performance of our incentive mechanisms.