MPS: A Truth Discovery Service Scheme by Using History Data to Maximize Profit for Mobile Crowd Sensing

Abstract

Mobile Crowd Sensing (MCS) has emerged as a novel paradigm in massive data collection, which leverages many individual mobile devices (called workers) to collect data. MCS platform utilizes the collected data to construct various services for service requesters, thus obtaining profit based on the data values contributed by workers. However, untrustworthy data would greatly reduce the data value, leading to a decline in platform profit, so it is crucial for the platform to recruit high-trust workers and collect truthful data, thereby providing high-quality service and obtaining high profit. To address this problem, we propose a Maximize Profit Scheme, called MPS, for MCS platforms, which consider that the data value declines as data trust decreases and discounts over time. MPS scheme is the first work that systematically addresses the impact of untruthful data on the platform profit, which is not well addressed in previous research. First, we utilize historical data of trusted workers as truthful data to identify the truth of data, which is a low-cost method. Then, a trust-discounting and time-discounting value model is proposed, which is more practical than previous methods. Based on the proposed value model, we propose a novel worker recruitment strategy combined with a trust-related and time-dependent reward threshold, which prioritizes workers with high trust and low latency, thereby promoting the data value of workers and maximizing the platform's profit. By comparing the MPS with existing schemes, the experimental results show that our MPS can achieve better performance in terms of total profit.