An Investigation of LPI Radar Waveforms Classification in RoF Channels

Turki Alrubeaan,Saeed Aldosari,Saleh Alshebeili,Amr Ragheb,Majid Altamimi,Khalid Albagami

doi:10.1109/access.2019.2938317

Turki Alrubeaan, Saeed Aldosari + Show 4 more

Open Access

https://doi.org/10.1109/access.2019.2938317

Copy DOI

Abstract

Intensive research has been developed to either design or classify low probability of intercept (LPI) radar signals. These types of signals are used in different sensitive electronic warfare applications such as electronic support, electronic attack, and radar emitter identification. Linear frequency modulation, nonlinear frequency modulation, frequency shift keying, polyphase Barker, polyphase P1, P2, P3, P4 and Frank codes are examples of LPI waveforms. In this paper, we consider the modulation classification problem under the effect of transporting the captured radar signals through radio over fiber channels. Distortions and noise introduced by such channels are likely to affect the performance of LPI classification algorithms. Here, we investigate the accuracy of a recently proposed hierarchical decision-tree automatic modulation classification algorithm for additive white Gaussian noise channels and provide the necessary adjustments when the intercepted radar signals are transmitted over fiber optic channels. The investigation is conducted by simulations and experimental demonstration. The obtained results show that for an 80 km fiber link and noisy intercepted LPI signals, the average identification accuracy reaches more than 98%, at 16 dB optical signal-to-noise ratio.

Highlights

In today’s battlefields, most radars, as in surveillance, reconnaissance, and target tracking radars, have to cope with very capable and advanced threats designed to contribute to the degradation of radar performance [1], [2]
This paper aims to study the optical fiber transmission medium and its effect on the intercepted low probability of intercept (LPI) radar signals
Both simulations and experiments have been considered for evaluating the performance of a recently proposed automatic classification algorithm in the presence of additive white Gaussian noise (AWGN), chromatic dispersion (CD), and amplified spontaneous emission (ASE) noise

Summary

INTRODUCTION

In today’s battlefields, most radars, as in surveillance, reconnaissance, and target tracking radars, have to cope with very capable and advanced threats designed to contribute to the degradation of radar performance [1], [2]. Features extracted from the signal directly (e.g. second order statistics, power spectral density, etc.) and those obtained from the 2D image of a time-frequency distribution are applied to Elman neural network (ENN) This network has a feedback, its recognition rate is relatively high at low SNR. We study the performance of the recently published AMC technique presented in [11] when the captured LPI radar signals are transported through fiber optic channels, as shown in Fig.. This AMC technique is based on well-known transform operations and a simple thresholding technique.

LPI RADAR WAVEFORMS CLASSIFICATION

Findings

CONCLUSION