I. Introduction
Electronic intelligence and electronic support measure (ESM) systems are passive electronic warfare (EW) systems that analyze the received radar signals without any transmission [1]. The detected pulses are encoded into pulse description words (PDWs), which contain interpulse parameters, such as time of arrival (TOA), pulse width (PW), frequency, and angle of arrival (AOA). Note that, at this point, each PDW can belong to different radar, since they are interleaved in the time domain. To identify the radars from the collected PDWs, they must be deinterleaved first [2]. When the signals coming from different emitters are well-separated in PW, frequency, or AOA domains, basic clustering algorithms might successfully deinterleave the collected PDWs without the TOA analysis [3]–[5]. However, as a result of the rapid developments in the fields of communication and radar systems, the frequency spectrum for a typical EW environment is getting more crowded, hence simple clustering techniques become insufficient for successful deinterleaving [6]–[8]. In addition, many EW systems are not able to measure AOA for each received pulse, and AOA information may be missing in the generated PDWs, which also affects the clustering performance adversely. Therefore, there is a need for a deinterleaving algorithm, which is capable of the TOA analysis. Our work tackles the scenarios, in which the radar signals have indistinguishable or unavailable PW, frequency, and AOA parameters, and it presents a novel TOA analysis algorithm.