Multi-Tasking for Cost-Efficient Mobile Crowdsensing under Uniformity Constraints

Abstract

In a practical Mobile Crowd Sensing (MCS) system, there usually exist multiple heterogeneous MCS tasks, each with a different set of requirements for completion. The coexistence of these MCS tasks presents a need to study the user selection and task allocation problem while considering factors like task and user heterogeneity, coverage, sensing data quality and total cost. In this paper, we study this issue by formulating a Multi-task User Selection (MTUS) problem with the aim of minimizing the total number of recruited workers subject to task requirements, user sensing capability while maintaining coverage uniformity. We show that our formulated problem is NP-hard. Consequently, we propose two variants of a greedy heuristic where the decision criteria for recruiting users is based on the sensing capability and the coverage contribution to the final workers set. A simple cost-efficient incentive scheme that reduces costs for task creators and increases profitability for task workers is also proposed. We perform simulations to test our model and we show that it achieves high coverage uniformity while reducing the number of users compared to a single-task oriented scheme.