PPHMA: Privacy-Preserving Hybrid Multi-Task Allocation for Mobile Crowd Sensing

Abstract

With the widespread adoption of mobile smart devices, mobile crowd sensing(MCS) has provided better services for people. To meet the growing sensing demands within a limited budget, platforms have integrated two modes—opportunistic sensing and participatory sensing—to utilize their complementary strengths. However, location privacy issues may reduce workers' willingness to participate, thereby affecting task completion rates. Although existing methods have addressed privacy protection in a single sensing mode, there remains little focus on location privacy in hybrid sensing modes. There are two main limitations in privacy issues related to task allocation: (i) how to effectively preserve workers' location privacy in hybrid sensing modes, and (ii) the usual reliance on trusted third-party institution. To address these issues, we propose a privacy-preserving hybrid multi-task allocation for MCS (PPHMA). This approach preserves workers' location privacy without relying on a fully trusted third-party institution, while maximizing the number of tasks completed. Specifically, for opportunistic task allocation, we employ zero-knowledge range proofs to protect workers' location, thereby avoiding location privacy leaks. Subsequently, based on the performance capability indicator of opportunistic workers, we select appropriate workers for task allocation. For participatory task allocation, we employ a worker location obfuscation generation algorithm to locally generate and upload obfuscated locations, ensuring that both the worker's real and obfuscated locations satisfy $\varepsilon$-Geo-Indistinguishability within the protected range. Then, based on the execution capability indicator of the participatory workers, we screen for candidate workers and use a greedy immune clone algorithm to optimize the workers' travel distances. Finally, we verify the effectiveness of the scheme through experiments using two real-world datasets.