A DL-Based Mutual Interference Suppression Approach for Uncoordinated Radar and Communication Coexistence System

The increasing demand for spectrum resources and their shortage make it inevitable for radar to share spectrum with communications, particularly for intelligent transportation systems (ITS). However, the coexistence of communication and radar leads to mutual interference, especially when the two systems are uncoordinated. Furthermore, complex and dynamic road environments limit the effectiveness of parameter estimation-based interference suppression methods. In this paper, we propose a deep learning (DL)-based interference suppression method, which can effectively suppress radar interference under uncoordinated scenarios. The received signal is first transformed into timefrequency domain by short-time Fourier transform (STFT). The magnitude of time-frequency representation is then fed into the network. After feature extraction and decoding, the estimated interference signal is obtained as the net output. The interferencefree signal is subsequently reconstructed by subtracting the estimated interference from received signal. Simulation results validate the effectiveness of proposed algorithm.