Corner Reflector Identification Based on Improved Temporal Convolutional Network and LSTM

The corner reflector is a typical passive jamming. To address the problems of relying on manual feature extraction and low accuracy of small samples in existing corner reflector identification methods, this paper proposes a method based on a combination of improved temporal convolutional network (TCN) and long short-term memory (LSTM). The proposed method uses TCN to extract deeper features of high-resolution range profile (HRRP) sequences, replaces the original cross-layer connection of TCN with dense connection, and utilizes LSTM to extract long-distance dependence information. Experimental results show that the identification accuracy of this fusion network reaches 99.12%, and the accuracy and robustness are better than the original model.