Strategies for information injection for networks estimating emitter location

Abstract

This paper considers the problem of emitter location in the presence of an adversary able to inject false information, where both the viewpoints of the adversary and locating network are considered. We begin by developing adversary strategies, which focus on degrading and redirecting the locating network. The locating network uses the time difference of arrival (TDOA) method and estimates the emitter location using both nonlinear least squares (LS) and the robust least median squares (LMS). Specifically, we determine the false sensor position the adversary should inject that maximally impacts the locating network by 1) minimizing the network’s accuracy, and 2) redirecting the network’s emitter location estimate. Similarly, we formulate a network’s strategy for detecting the adversary’s injection and develop a detector able to determine its presence under both adversary strategies. We present both theoretical and numerical results illustrating the effectiveness of the developed strategies.