A Secure Auction Mechanism for Task Allocation in Mobile Crowdsensing

Mobile crowdsensing has attracted widely attention as a new sensing paradigm, in which mobile users collect sensing data by their devices embedded various sensors. To motivate mobile users participating in sensing tasks, a number of auction mechanisms have been proposed. In our work, we focus on the task allocation problem with multiple constraints for the auction-based crowdsensing system to maximize profit of the central platform, which has been proved to be NP-hard. To solve the problem, a greedy-based task allocation algorithm with $$(1+\gamma )$$ -approximation solution is proposed, in which the bid improving profit of the platform most is selected as the winning bid greedily in each iteration. However, bids for all tasks of a user submitted to the platform might let out location of the user unexpectedly. Therefore, we further design a secure auction mechanism with secret-sharing-based task allocation protocol, where each user can submit at most a winning bid to the platform instead of all bids for tasks to prevent locations of users from being inferred. The effectiveness of task allocation and location privacy protection based on our proposed secure auction mechanism is verified by theoretical analysis and simulations.