Protecting Bilateral Location Privacy in Task Allocation With Differential Privacy in Crowdsensing

To address the critical issue of user location privacy and enhance the security of user information, the Laplace mechanism, a differential privacy technique, is frequently utilized as a privacy protection method for safeguarding user location privacy. Nevertheless, the direct application of Laplace noise to location data faces challenges in achieving a balance between selection accuracy and privacy preservation. Furthermore, previous studies have largely overlooked the privacy considerations associated with task locations when safeguarding user location data. To address these challenges, this paper introduces a novel Differential Privacy-based Bilateral Location Protection Scheme (BLPTAC) specifically designed for task allocation in crowdsensing. Our innovative approach transforms users' two-dimensional coordinates into one-dimensional indexes using the Hilbert map, followed by the local application of Laplace noise to generate a disturbance index. This process enables the server to employ mathematical integration for distance comparison of the disturbance index. This allows our scheme to not only enhance privacy protection beyond conventional differential privacy mechanisms but also boost selection accuracy. Furthermore, to secure the location information of tasks, our method incorporates density clustering and group signature techniques. The efficacy of our proposed scheme is rigorously assessed using public datasets, including Gowalla and T-driver, as well as a proprietary dataset, OurCheckIn. The experimental results demonstrate that our approach exhibits superior performance in achieving remarkable selection accuracy while robustly protecting location privacy.