Privacy preserving RSS map generation for a crowdsensing network

Abstract

Nowadays advanced mobile computing needs accurate RSS maps for effective AP deployment and mobile applications. Inspiringly, the emerging crowdsourcing paradigms provide an innovative and effective way for large-scale RSS gathering. However, existing methods need sampling data and location information from participants, which could be a serious threat to privacy. In dealing with this difficulty, we present a privacy preserving RSS map generation scheme for crowdsensing networks called PRESM. To protect the privacy of user traces, we exploit the compressive sensing technique to sample and compress RSS values along each road segment, which removes the temporal and concrete location information of each participant. Meanwhile, each smartphone user carefully selects a subset of road segments to send its compressed RSS data to a third party. The third party component provides better privacy protection by removing more road segments, and the central server is responsible for RSS map generation. Finally, we carry out our experiment on a campus of approximately 1.6 km2 . Experimental results demonstrate that an RSS map is generated relatively accurately without sacrificing users' trace privacy, and the coverage ratio of the geographic map is greater than 90 percent.