| This thesis study and address: (i) the location spoofing attack in time-of-arrival (TOA)-based localization systems, (ii) the traffic analysis attack in wireless networks, and (iii) privacy-awareness in mobile crowd sensing (MCS) applications. First, we study how to detect location spoofing attacks in TOA-based localization systems and design a generalized likelihood ratio test (GLRT) to detect location spoofing anomalies in the received TOA delay measurements. Second, we study how to provide privacy for a communicating source-destination pair and design a statistical decision-making framework to select the routing path distribution that minimizes the detection probability of Bayesian maximum-a-posteriori (MAP) inference (a type of traffic analysis method). We then formulate linear programs to minimize the expected detection probability of a MAP adversary, subjected to a privacy budget constraint. We also propose the (k, e)-anonymity privacy constraint for strict privacy guarantees in wireless networks and formulate a mixed-integer linear program to minimize the expected routing cost of the (k, e)-anonymous path distribution. Finally, we study how to improve the utility (in terms of spatial coverage) of privacy-aware mobile crowd sensing applications and propose a privacy-aware Stackelberg game incentive mechanism to select the optimal set of participating mobile smartphone users. |
