Spatial Stackelberg Incentive Mechanism for Privacy-Aware Mobile Crowd Sensing
Journal of Machine Learning Research 1 (2000) 1-48
31 Pages Posted: 9 May 2018
Date Written: 2000
Mobile crowd sensing is an emerging sensing paradigm where sensing applications buy sensor data from spatially distributed mobile smartphone users (workers) instead of deploying their own sensor networks. This reduces costs and enhances the range of possible applications that can be developed, since the application developers can focus on other tasks such as spatial field reconstruction or estimation of some spatial characteristics for a process being sensed.
In many spatial monitoring applications, the crowdsourcer needs to incentivize smartphone users to contribute sensing data such that the collected dataset has good spatial coverage and can be used for accurate spatial regression that will meet quality of service requirements applicable to the product on offer from the application.
To further incentivize privacy-concerned workers to contribute, we propose a novel Stackelberg incentive mechanism that allows workers to specify their location whilst satisfying their location privacy requirements. We then derive a unique Stackelberg equilibrium which demonstrates the stability of our approach. Next, we prove the existence of a Stackelberg equilibrium in this spatial game context in which the crowdsourcer imposes constraints on the minimum and maximum data contributions for each user, and we study sufficient conditions for achieving Pareto efficiency. We develop efficient algorithms to solve for the equilibrium for the follower and leader games via a backward induction formulation.
We apply our theoretical and methodological results to a real-world application and our simulation results show that our proposed Stackelberg incentive model is better in terms of predictive mean and variance compared to the disk and k-depth coverage maximizing schemes.
Suggested Citation: Suggested Citation