Research on Radar Signal Sorting Method Based on Improved Adaptive Density Clustering

To address the challenges of low accuracy in radar signal sorting under complex electromagnetic environments, the difficulty in setting priori parameters for DBSCAN clustering algorithms, and poor performance in sorting non-uniform, multi-density radiation source signals, this paper proposes an improved adaptive density clustering algorithm. This method achieves sorting of unknown radar radiation sources with minimal reliance on priori parameters. Leveraging the strength of the Density Peak Clustering (DPC) algorithm in effectively identifying outliers, it removes noise points. Subsequently, a nearest neighbor graph is constructed to identify local high-density points based on its inherent adaptive density properties. Finally, clusters are expanded according to density connectivity principles to complete the clustering process. Experimental results demonstrate that the proposed algorithm effectively resolves the parameter configuration challenge of DBSCAN density clustering and achieves accurate clustering for non-uniform density emitters, proving particularly suitable for real-time sorting of unknown radar emitters in complex battlefield electromagnetic environments.