DP-Loc: A Differential Privacy-Based Indoor Localization Scheme with Bilateral Privacy Protection

Among indoor localization services, WiFi fingerprint technology has attracted much attention because of its wide coverage area and high positioning accuracy. However, concerns arise regarding the vulnerability of users’ personal sensitive information to malicious attacks during location queries. Existing differential privacy (DP) indoor localization techniques encounter challenges when iterating the collected user fingerprints into the clustering process, as the addition of noise to cluster centroids during iterative operations can lead to increasing deviations, adversely affecting clustering results and causing localization errors. In this paper, we put forward a differential privacy indoor localization scheme (DP-Loc) that supports bilateral privacy protection. It mainly protects the privacy of the user’s personal location when querying in the online localization phase, as well as the privacy of the fingerprint database on the server side. The online stage server-side fingerprint clustering reduces the error in the fingerprint clustering stage by selecting the centroids, dividing the similarity between clusters and clusters, adding noise to cluster centers based on differential privacy, masking the true position by fingerprint replacement, and returning the results to the client to complete the localization. Experimental simulations and tests conducted on an existing dataset demonstrate that the DP-Loc scheme enhances localization accuracy within the same privacy budget while also providing more robust protection for user location privacy compared to existing approaches.