An Analysis of Process Parameters for the Optimization of Specific Emitter Identification Under Rayleigh Fading

Abstract

Specific Emitter Identification (SEI) was introduced over twenty-five years ago to grant electronic warfare systems the ability to uniquely, distinguish between RADARs of the same type using intra-pulse modulation features. The demonstrated success of RADAR SEI led to its application in the identification of wireless transmitters for the purpose of augmenting digital security measures within public and private communication networks. The majority of SEI work has focused on wireless channels comprised of noise only, unknown multipath fading conditions, or fixed fading channels. Our recent works focused on SEI performance within multipath fading channels that conform to a known model/behavior and change from one transmission to another. However, these works did not design the SEI process (i.e., from waveform collection to final radio identification decision) with the singular purpose of maximizing SEI performance under multipath fading. In order to maximize SEI performance under Rayleigh fading, this work analyzes the impacts of the: (i) filter type, order, and bandwidth; (ii) Gabor Transform analysis window width; as well as (iii) the number of candidate signals used by the Nelder-Mead (N-M) channel estimator. The result is a 11.9% average percent correct classification performance improvement for a length five Rayleigh fading channel at a signal-to-noise ratio of 9 dB.