CNN-LSTM Radar Signal Sorting Algorithm Fusing Intra-pulse and Inter-pulse Multi-dimensional Features

Aiming at the low sorting accuracy of the traditional radar signal sorting algorithm under the aliasing of parameters and modulation forms of radar signal pulses, this paper propose a radar signal sorting algorithm based on the fusion of intra-pulse and inter-pulse multi-dimensional features. Firstly, multiple intra-pulse features of radar signals are extracted and combined with inter-pulse parameters to form a multi-dimensional feature matrix. Then, signals are pre-sorted through the DBSCAN clustering algorithm. Finally, the main sorting of radar signals is realized based on the proposed CNN-LSTM neural network. Simulation results demonstrate that compared with support vector machine (SVM), CNN, and LSTM networks, the proposed CNN-LSTM algorithm has great advantages in the sorting accuracy of radar signals, reaching 98.3%.