Time-dependent distributed collaboration and incentive mechanism for Mobile Crowdsensing

Abstract

In Mobile Crowd Sensing (MCS), the increasing complexity of sensing tasks and the rising demand for data quality have rendered traditional single-participant sensing paradigms inadequate. Collaborative sensing involving multiple participants has emerged as a crucial approach to enhance sensing efficiency and accuracy. However, centralized collaboration strategies often impose significant computational and processing burdens on the platform, while neglecting participants’ actual capabilities and willingness to cooperate. Moreover, existing research rarely addresses the sensing time redundancy that arises when multiple participants collaborate on the same task. Additionally, most studies assume participants have long, continuous time slots available for sensing, which does not align with real-world scenarios where participants’ available time is often fragmented. To address these challenges, this paper proposes a Time-dependent Mobile Crowdsensing Distributed Group Collaboration System (TMDCS). First, we construct a task selection model that accounts for participants’ sensing capabilities and allocates different suitable tasks across their multiple fragmented time slots. We also develop a task collaboration incentive model aimed at encouraging greater participation and ensuring high-quality sensing data. Second, a distributed task optimization mechanism is designed to improve overall social welfare. This mechanism selects leaders and forms collaborative groups based on coalition game theory. Finally, a reverse auction scheme is applied to select the optimal coalition for each task and determine incentive distribution. Experimental results demonstrate that the proposed TMDCS outperforms baseline methods, achieving average improvements of 49.7% in social welfare, 37.5% in task coverage, and 25.3% in task quality.