Characterizing Non-Linear Modulations on radar pulse under non-cooperative estimations

Modulation recognition is an important aspect of modern Electronic Support Measure (ESM) systems for effective analysis of intercepted radar pulses. Pulse compression radars use a variety of complex non-linear modulations to have better side lobe rejections as compared to linear modulation. This paper investigates numerous techniques of non-linear frequency variations in radar waveforms and proposes an algorithm for their passive detection and classification. The proposed methodology uses instantaneous phase dependencies as a statistical metric for modulation categorization. The non-linearity once detected, is further classified as sinusoidal or tangential waveform. The performance of the proposed algorithm is evaluated through Monte-Carlo simulations against a variety of input configurations. It successfully achieves the detection probability of above 90% at moderate SNRs. The practicality of the scheme has also been discussed regarding its time complexity for real time applications. It is concluded through simulations that if enough data length is selected to catch the non-linear curve, the proposed methodology achieves better real time detection and estimation performance than the BFD, CFD and kay-windowed phase detection techniques.