| This paper mainly studies the technologies of the radar emitter pulse modulation recognition, respectively researching the recognition algorithm based on time-frequency characteristics and the recognition algorithm based on neural network.At first, this paper, using the correlation detection algorithm, accurately estimates the time of radar emitter signal arrival. Phase unwrapping algorithm is studied and the real phase of radar signal is deduced from the measuring phase of signal. Based on the instantaneous phase of signal, the instantaneous frequency of the signal is obtained with the improved IKay algorithm and the IKay algorithm performance is analyzed.Based on the instantaneous phase and instantaneous frequency of the signal, this paper analyzes the twelve kinds of radar emitter signal model and expounds the characteristics of each signal in time domain and frequency domain. Also the time-phase and time-frequency features of the different signals are compared to summarize their identifical features of intra-pulse modulation. Based on signal time and frequency characteristics, studied modulation recognition and parameter estimation method of the 12 kinds of commonly used radar signals are given. By computer simulations, the recognition rate and parameters estimation error are figured out, which meet the engineering requirements.Finally, this paper studies a kind of new characteristic parameters of radar pulse signal. The characteristic parameter mainly refers to the statistical parameter of signal amplitude and similarity feature of signal spectrum. A hierarchical structure of BP neural network classifier is designed. Based on this classifier, the experiment on recognition of six kinds of radar emitter signals is carried out. And the simulation results show that the algorithm has a high correct recognition rate. |
PDF Full Text Request