A Novel Radar Emitter Fingerprint for Multiple Path Propagation Environment

Mining additional modulation information of signal propagation environment as a new fingerprint feature is more important for modern radar reconnaissance systems. This paper proposes a definition and processing method of a new radar emitter fingerprint from multiple path propagation environment, and demonstrates its application in an improved radar pulses deinterleaving task. For the fingerprint generation, the proposed method reconstructs a reference radar pulse by pulse description words (PDW) and correlate it to intercepted pulse sequence firstly. Then an adaptive threshold is utilized to pick out the main peaks belonging to the same radar. Next all the selected range profiles take a procedure of self-focusing and coherent accumulation among different pulses. Finally, propagation environment fingerprint (PEFP) is obtained after eliminating main-peak by CLEAN in the accumulated results. For the deinterleaving application, a sparse auto-encoder (AE) is used to extract high dimensional features from the generated PEFPs, then the features are combined into traditional PDW as extended features to implement an improved deinterleaving task. Simulation results verify the validity and effectiveness of the proposed fingerprint. The PEFP enhanced deinterleaving processing can improve the precision by 6% in non-cooperative conditions compared with traditional methods.