Multifunction radar signal sorting method based on data field and Finch-Means

In addressing the issues faced by traditional signal sorting methods and clustering algorithms in complex electromagnetic environments, such as the need for prior knowledge, initial center selection, and signal batch increment, a multifunctional radar signal sorting method based on Data Field and Finch-Means is proposed. Initially, the integration of the signal envelope "pulse leading edge" in the radar pulse signal fingerprint features into the Pulse Descriptor Word (PDW) is proposed, forming a new feature parameter (EPDW) used by the Data Field for the elimination of interference pulses. Concurrently, a K-means clustering algorithm based on the first-neighbor points (Finch-Means) is introduced. This algorithm employs the Finch algorithm to calculate the first neighbor vectors of the filtered pulses, preclassify the data, determine the number of clusters, and identify the initial cluster centers. Finally, the K-Means algorithm computes the distances from the pulses to each cluster center and assigns them to the nearest cluster center, completing the final sorting. Simulation experiments show that the average recognition rate of this algorithm reaches 97.85%. Compared to existing methods, the average correct recognition rate of a single radar improves by 14.88%, and the average recognition rate increases by 16.55%.