Employing Social Participants for Timely Data Collection Using Pub/sub Solutions in Dynamic IoT Systems

Abstract

In Mobile Crowd Sensing (MCS), the platform publishes sensing tasks with the requirements of location, time and data attributes and hires participants to perform sensing tasks, thus it can collect and process mass of sensing data to construct various applications and publish them to service requesters, which is a low cost and effective method for large-scale data processing. However, most of the existing studies do not consider the value of tasks with time-discount property and insufficient participation, which results in delayed task completion and low task completed ratio. In this paper, we propose data collection mechanisms in offline and online scenario considering both of these factors to collect data timely and increase the benefit of the platform effectively. We propose a multi-tiered spreading task structure, in which participants act as agents to recruit their social neighbors to perform tasks, and the social neighbors can be recruited as new agents to spread tasks so that there is a sufficient number of participants in the platform to be selected to perform tasks. In winner and agent selection of online mechanism, we change the allocation of budget in multi-stage sampling accepting process based on real-time task completed ratio. We proved that the proposed mechanisms achieve computational efficiency, truthfulness, individual rationality and budget feasibility, and after extensive experiments, we proved the proposed mechanisms are superior to previous strategies.