Privacy-preserving Batch-Based Task Assignment over Spatial Crowdsourcing Platforms

With the rapid development of mobile networks and the widespread use of mobile devices, an increasing number of spatial crowdsourcing platforms have emerged. Task assignment is a crucial aspect of spatial crowdsourcing, where workers must provide true location information to the server to ensure matching with the nearest tasks. However, it will result in the leakage of location privacy for workers and tasks. In light of this, existing works utilize methods based on differential privacy or homomorphic encryption, which suffer from issues of low service quality and high computational costs. In this paper, we propose a two-stage location privacy protection framework that balances service quality and computational efficiency. The framework initially divides workers and tasks into groups based on location relevance and utilizes secure computation with homomorphic encryption to enable task sharing within each group, improving task assignment. We propose a distance evaluation method based on Mahalanobis distance to measure the correlation between workers and tasks in perturbed locations. Moreover, by designing a K-Means-based grouping algorithm, we cluster workers who can effectively share tasks, increasing the success rate of task assignment through task swapping. Finally, we verify the effectiveness and efficiency of our method through synthetic and real datasets.