Radar Signal Sorting Technology In Complicated Environments

Radar signal sorting is to distinguish each radar's emission signal from their aliasing signals. In the current electronic countermeasures wax , it's extremely important to obtain information about the enemy accurately.The signal sorting is the key technology of electronic reconnaissance system. In today's environment,electronic environment is complex, signal overlapping is serious, and there are also many unknown radar radiation source which can make interferences. Therefore,how to separate radar signal from others is the most urgent problem that to be solved.This paper firstly makes a brief study on the K-means clustering algorithm and Fuzzy clustering algorithm: the K-means clustering algorithm is Hard classification of radar samples,whose clustering accuracy is not high. In contrast, Fuzzy clustering algorithm is need to set a prior information., which cannot cluster effectively unknown radar emitter signal. In view of weaknesses of the traditional clustering algorithm, firstly, beginning from the interstitial features of radar signal and using the five traditional parameters, this paper presents a modified FCM algorithm based on invasive weed. Then, it researches the method of extracting the features of radar signal in a pulse, and proposes a method for extracting interstitial features based on time-frequency atoms. The main work and achievements are as follows:For the characteristics of the interstitial cluster of radar signals, Invasive weed algorithm has a lot of advantages:Its structure is simple. Its parameters are few,and its capability of global search is also better. Besides, it can get the optimal solution in less number of iterations. For that, this paper proposes a modified FCM algorithm based on invasive weed. This algorithm primarily ameliorate Fuzzy clustering algorithm's dependence for initializing cluster centers. First, according to the number of samples, determining the solution space the number of radars' categories. Then on the basis of the distance criterion,using the weed algorithm search the optimal number of categories in the whole solution space,as the initial parameters of the fuzzy clustering to cluster. Compared to the K mean clustering algorithm and Fuzzy clustering algorithm, it proofs that the algorithm gets rid of the dependence on initial clustering center and has a better sorting accuracy.For the clustering of the interstitial features of radar signal, using the interstitial features of radar signal to research the signal sorting is the hot spot currently. Many researchers proofed that extracting interstitial features based on time-frequency atoms is effective.However, the number of time - frequency atoms is huge. Its computational complexity is high.For solving this problem, this paper proposes a modified method of extracting interstitial features based on time-frequency atoms. First, it introduces five traditional classic mathematical model of Radar signal.And then, it proposes the method of the combination of the weed algorithm and time-frequency atom. According to the distance criterion, A group of atoms with Searching by weeds intelligent algorithm,we can distinguish different atoms of different modulation modes of radar signal, and make the inner product operation wiht the radar signal to be sorted, which is the classification of the input vector of the improved FCM algorithm. Making simulation experiments under the SNR from -5dB to 3 dB proves the effectiveness of the algorithm.