A Nonparametric Method for Detecting Unintended Electromagnetic Emissions

Unintended electromagnetic emissions of RF transceivers are signals that are emitted by all RF devices and often lie within the noise band. The identification of such signals becomes an important issue as the identification, localization, and recognition of malicious wireless devices could result in a passive means to render such devices useless. In this paper, we present a new nonparametric method to distinguish signals from noise based in a morphological trimming of the data complemented with an analysis of the shape of the sequence of curves in which the data series can be decomposed. The method takes its tools from the mathematical morphology and multivariate statistics. The good performance of this methodology is illustrated with a set of real data coming from three small RF transceivers such as two-way talk radios, which are commonly used in improvised explosive devices.