Sensor source location privacy based on random perturbations

Sensor source location privacy, which means to protect source sensors' locations of network traffic, is an emerging topic in wireless sensor networks, because it cannot be fully addressed by traditional cryptographic mechanisms, such as encryption and authentication. Current source location privacy schemes, assuming either a local attack model or a global attack model, have limitations. For example, schemes under a global attack model are subject to a so called `01' attack. Targeting on solving this attack under a global attack model, we propose two perturbation schemes, one based on Uniform distribution and the other based on Gaussian distribution. We analyze the security properties of these two schemes. We also simulate them and compare them with previous schemes, with the results showing that the proposed perturbation schemes can improve the source location privacy significantly.