A new clustering and sorting algorithm for radar emitter signals

Abstract

The traditional k-means clustering algorithm needs to set parameters manually in radar signal sorting application, which is sensitive to isolated points and prone to “batch” phenomenon, and the selection of initial clustering center has a direct impact on clustering effect. In order to solve the above problems, a radar signal sorting algorithm based on data field and improved k-means clustering is proposed. Firstly, the potential values of all data samples are calculated according to the theory of data field. After the isolated points are removed, the maximum local potential values are found. The nearest sample data is selected as the initial clustering center, and the maximum number of local potential values is selected as the clustering number. Finally, the improved k-means clustering algorithm is used to complete the radar signal sorting. The algorithm can automatically obtain the initial clustering center and the number of clusters, and the clustering results have larger inter cluster distance and smaller intra cluster distance, so the clustering results are better. Simulation results verify the feasibility of the algorithm.