Workload-Aware Indoor Positioning Data Collection via Local Differential Privacy

The rapid development of indoor localization technologies has made it easier than ever for service providers to collect and analyze massive amount of indoor positioning data from users. However, an indiscriminately collecting indoor positioning data causes serious privacy problems because a user's sensitive information can be inferred by analyzing the indoor movement behavior. Therefore, in this letter, we propose a method to collect indoor positioning data from users by using local differential privacy (LDP) while preserving their privacy. Specifically, given an available target workload, the proposed method finds an optimal data encoding and perturbation scheme of LDP to minimize the overall estimation error corresponding to the workload. Experimental results using synthetic and real datasets confirm that the proposed workload-aware scheme can achieve better accuracy than the workload-independent methods.