Recognition of Micro-Motion Jamming Based on Complex-Valued Convolutional Neural Network

Abstract

Micro-motion jamming is a new jamming method to inverse synthetic aperture radar (ISAR) in recent years. Compared with traditional jamming methods, it is more flexible and controllable, and is a great threat to ISAR. The prerequisite of taking relevant anti-jamming measures is to recognize the patterns of micro-motion jamming. In this paper, a method of micro-motion jamming pattern recognition based on complex-valued convolutional neural network (CV-CNN) is proposed. The micro-motion jamming echo signals are serialized and input to the network, and the result of recognition is output. Compared with real-valued convolutional neural network (RV-CNN), it can be found that the proposed method has a higher recognition accuracy rate. Additionally, the recognition accuracy rate is analyzed with different signal-to-noise ratio (SNR) and number of training samples. Simulation results prove the effectiveness of the proposed recognition method.