Allocating Heterogeneous Tasks in Participatory Sensing with Diverse Participant-Side Factors
Jiangtao Wang,Feng Wang,Yasha Wang,Daqing Zhang,Brian Lim,Leye Wang
DOI: https://doi.org/10.1109/tmc.2018.2869387
IF: 6.075
2019-01-01
IEEE Transactions on Mobile Computing
Abstract:This paper proposes a novel task allocation framework, PSTasker, for participatory sensing (PS), which aims to maximize the overall system utility on PS platform by coordinating the allocation of multiple tasks. While existing studies mainly optimize the task allocation from the perspective of the task organizer (e.g., maximizing coverage or minimizing incentive cost), PSTasker further considers diverse factors on the participants' side, including user work bandwidth, user availability, devices' sensor configuration, task completion likelihood, and mobility pattern. Furthermore, by considering the heterogeneity in three dimensions (i.e., task, time, and space), it adopts a novel model to measure task sensing quality and overall system utility. In PSTasker, it first calculates the utlity of a given task allocation plan by jointly fusing different participant-side factors into one unified estimation function, and then employs an iterative greedy process to optimize the task allocation. Extensive evaluations based on real-world mobility traces demonstrate that PSTasker outperforms the baseline methods under various settings.