Providing location and location pattern privacy on location-based applications

Location privacy is becoming an important issue with the advances in mobile technology and widespread use of location based applications. Hence, it is extensively studied in the context of Location Based Services (LBSs). In a typical solution, users are assigned a location privacy profile and the precise locations are cloaked so that the location privacy is not compromised. The process is well-defined for snapshot location privacy. On the other hand, such solutions require patches in case of multiple LBS requests on the user trajectory. The fact is that the attacker can exploit some background knowledge including maximum velocity to exactly locate the user. PROBE and other systems operate in this fashion to detect any privacy violations and apply necessary time-delaying or post-dating. Taking the data mining as a stronger adversary, user-specific private patterns can be extracted from the respective trajectory history. This thesis defines location pattern privacy problem as a complementary to location privacy problem. As a result, a framework in which user-specific sensitive location patterns are sanitized online fashion is introduced. The solution uses an efficient dynamic programming approach to check pattern violations in incremental manner. The tool implementing this solution and an experimental evaluation are presented as well.