Research On Unknown Emitter Signal Sorting Algorithm Under Complex Signal Environment

With the continuous development of radar signal modulation technology, the work system of emitter changing quickly, the signal sorting environment becomes increasingly complex. So,the difficulty of signal sorting continue to increase. This paper will expand researches on the signals sorting of unknown emitter under complicated environment.Noticing that the sorting effect of a single algorithm is not ideal , we will study sorting algorithm based on the combination of the different types of algorithms.First, analyze the characteristics of signals from the frequency domain, time domain and spatial domain parameters respectively. Focus on the time domain parameters and the main way of modulation. Detailed introduces several different modulation types of emitter signals and use MATLAB to the emulate signal emitters. Construct the signal sorting model. Briefly describes the working process of traditional PRI algorithm. Focus on introducing PRI transform. Solving the problem of PRI transform under the large shaking and improve the algorithm. Then introduce the improved algorithm and do the simulation experiment.By observing the experimental results of every algorithm, determine their respective advantages and disadvantages.Second, briefly describes the Blind Source Separation and the basic working principle of ICA(Independent Component Analysis)algorithm,focus on the FastICA algorithm based on negative entropy maximization. The algorithm is applied to signal sorting fields. Experiments show that this algorithm is remarkable for sorting different kinds of modulation signal , but are sensitive to noise signal.Finally, for the signal sorting under complicated environment, a new algorithm based on sine wave extraction and improved pulse repetition interval(PRI) is proposed. Firstly, sine wave extraction is used to separate signals with different modulation types. Then, in the case of emitter pulse number decrease, improved PRI algorithm can be used to sort signals with the same modulation types but different modulating parameters further. Simulation result shows that the new method is accurate and effective.