Preserving location privacy against inference attacks in indoor positioning system

Location-based service (LBS) servers are refined periodically due to the new type of privacy issues involved while providing various social benefits to the users. In literature, many privacy mechanisms are proposed for LBS servers, but these mechanisms fail to resist many possible inference attacks, such as multi-time publication attacks and multi-time continuous publication attacks from untrusted LBS servers. Hence, the LBS providers cannot provide strict privacy guarantees to participating users. To this end, the LBS servers require a privacy mechanism under differentially private, which resists the above two inference attacks and one common inference attack i.e., one-time inference attacks. The proposed privacy mechanism involves two phases; the first phase is finding a set of obscure locations using a sampling technique, and the second is finding a suitable obscure location using a clustering technique. Finally, conduct a series of experiments in order to validate the effectiveness of the proposed privacy mechanism with other state-of-the-art methods. The experimental results evince that the proposed privacy mechanism resists three possible inference attacks and provides a strict privacy guarantee as in the traditional differential private mechanism.