Radar Emitter Signal Recognition Based On Time-Frequency ATOM Approach

As an important component of the ELectronic INTelligence (ELINT) systems, Electronic Support Measure (ESM) systems and Radar Warning Receiver (RWR) systems, radar emitter signal recognition, which provide precious information for early warning, threat detection, threat avoidance and counterstrike disposition, is a key link in the Electronic Warfare (EW) and determine the success or failure of the EW. Since the advanced radar with complicated systems has been widely used and played a major rule gradually in the filed of EW, the density, the operating frequency and the complexity of the radar signal have been increased significantly. For example, the density of the radar signal has reached million pulse in a second; the range of signal frequency has exceed GHz unit; and the modulation of interpulse and intrapulse become much more complex, especially the modulation parameters exist the cut-off change. Additionally, the enhancement of the Electronic CounterMeasures (ECM) and Electronic Counter-CounterMeasures (ECCM) makes the electromagnetic environment much more complexity. All those factors pose a serious challenge to the radar emitter signal recognition.Radar emitter signal recognition involves intentional modulation recognition and unintentional modulation recognition. The purpose of the intentional modulation recognition is to identify the modulated rule of the amplitude, phase and frequency, while the goal of the unintentional modulation recognition is to confirm the radar individual through the feature measurement. In recent years, both the intentional modulation recognition and unintentional modulation recognition of radar emitter signal have attracted extensive attention from the experts and scholars of the filed of EW. And varied recognition methods have been proposed, in which the interpulse-based recognition approaches are the most common techniques. In order to implement the intentional modulation recognition and unintentional modulation recognition, this dissertation makes an in-depth exploration and investigation of the interpulse feature of the radar emitter signal by the use of time-frequency approach. The main research work and research fruits are as follows:1,The investigation about the present situation of intentional modulation recognition and unintentional modulation recognition of the radar emitter signal, and the time-frequency atom approach is summarized and analyzed. The problems need to be solved are also pointed out. Besides the model of the time-frequency atom approach based radar emitter recognition is presented.2,A feature extraction method based on the Gabor and Chirplet time-frequency atom dictionaries is proposed to implement the intentional modulation recognition. Based on the characteristics of the intentional modulated radar emitter signal. Gabor and Chirplet time-frequency atom dictionaries are constructed, and an improved differential evolution based fast time-frequency atom approach is proposed to settle the issue of high computational complexity of the traditional time-frequency atom approach. Then five different typical radar emitter signals are decomposed in the dictionaries. By thorough analyzing the characteristics of the decomposed atoms, three time-frequency atom features are extracted for intentional modulation recognition. Experiment results show that under the circumstance of different modulation parameters and low signal to noise ratio, the proposed features can obtain a high correct recognition rate.3,By analyzing the origin and characteristic of the unintentional modulation of radar emitter signal, a phase noise model based on frequency source of the radar characteristic is constructed to simulate the unintentional modulation. Since the phase-lock loop frequency synthesizer with nice spectrum purity and high spurious suppression are wide adopted in modem radar transmitter, this dissertation analyzes the phase noise of differential devices in the phase-lock loop frequency synthesizer in detail, and then by the use of linear superposition principle, the integrated phase noise model of radar transmitter is presented. Additionally, through discussing the effect of phase noise on emitter signal, the simulation method of unintentional modulation radar emitter signal is presented, which solves the problem that the true radar emitter signal could hardly be obtained in the research procedure.4,A feature extraction method based on the Sinusoid time-frequency atom dictionary is proposed to implement the unintentional modulation recognition. Based on the characteristic of the unintentional modulation radar emitter signal with phase noisy, average accumulation of multi-pulse emitter is utilized for feature extraction to reduce the effect of the additive noise on phase noise. Then the Sinusoid time-frequency atom dictionary is constructed, and an membrane algorithm based fast time-frequency atom approach is proposed to implement the fast search of the best matched atom. By analyzing the characteristic of the conventional unintentional modulation radar emitter signal, two time-frequency atom features are extracted for unintentional modulation recognition. Experiment results demonstrate that in the situation of high signal to noise ratio, the proposed features can obtain a reliable preferable recognition rate.This work is supported by the National Natural Science Foundation of China (60702026) and the Scientific and Technological Funds for Young Scientists of Sichuan (09ZQ026-040).