SACD algorithm of pulse stream analysis

Final task of ESM (Electronic Support Measures) equipment is emitter identification. Identification in primary layer is based on processing of pulses stream existing on ESM equipment input. In this process each emitter is determined by set of parameters. Pulses in ESM equipment, during measurement process are associated with angle of arrival (AOA), frequency and time parameters. Those parameters constitute primary base for deinterleaving process. Time of arrival (TOA) is a basic parameter. It assures pulse repetition interval (PRI) estimation. PRI of radar signals usually is modulated in specific manner (e.g. stagger, sliding, dwell and switch) or in cyclic way and also it can have stochastic nature (e.g. jitter). Identification of PRI changes improves emitter description what has a considerable impact on recognition correctness. Detection of cycle in PRI modulation is relatively easy task from computation point of view and under assumption that acquired pulses stream can be associated with particular emitter. In practice due to antenna movement, ESM equipment acquires pulse trains (parts of pulses stream). Furthermore some pulses are missing and some additional pulses appear as result of noise and signal interaction. The methods of cycle detection presented in literature of problem have O(n²) computational complexity and require considerably big memory. The paper presents Sequential Algorithm of Cycle Detection (SACD) that has linear O(n) computational complexity. Results of examinations confirm high effectiveness of PRI parameters estimation for non-homogenous pulses stream.