Unsupervised k-means combined with SOFM structure adaptive radar signal sorting algorithm
Shunqi Su, Xiongjun Fu, Congxia Zhao, Jingfang Yang, Min Xie, Zhifeng Gao · 2019 IEEE International Conference on Signal, Information and Data Processing (ICSIDP) · 2019
With the overlapping of signal parameters, signal sorting faces great challenges. K-means clustering algorithm and self-organizing Feature Mapping (SOFM) neural network algorithm are widely used in radar signal sorting. However, the cluster number of k-means algorithm needs to be determined in advance, and the initial cluster center also needs to be randomly selected, so it is easy to fall into local optimal. The accuracy of SOFM neural network sorting results is greatly affected by the preset structure. Aiming at the above two problems, this paper introduces the density dynamic clustering into the traditional k-means clustering algorithm and combines it with SOFM neural network to put forward an unsupervised structural adaptive radar signal sorting algorithm. The simulation results show that the algorithm can effectively solve the problem of signal sorting in the case of parameter space overlap and the computation is small.