Location-Semantic Aware Privacy Protection Algorithms for Location-Based Services

摘要

Location-semantic has become one of the key factors that lead to privacy disclosure in LBS. To solve this problem, we first propose sensitivity-fading algorithm (SFA). As there may be various semantics on a road, we design the semantic-influence vector for each road in the road networks. By calculating the sensitivity of each road based on the semantic-influence vector and selecting the road with the lowest sensitivity, SFA generates the cloaking set (CS) quickly. However, it is vulnerable to inference attack and replay attack. Thus, we also propose the weight-based sensitivity-fading algorithm (WB-SFA). It calculates the weight of each candidate road according to the distance between the candidate road and the road where the user located and gets the sensitivity of each candidate road based on the road's weight and semantic-influence vector. The road with the lowest sensitivity would be added into the CS. The experimental results show that both the two algorithms can guarantee a high success rate in generating CS. The SFA can generate CS at a faster speed while the WB-SFA has strong capability for resisting the inference attack and replay attack.