Deep Learning-Based SNR Estimation with Covariance Input
Shurun Chen, Shilian Zheng, Zhuang Yang, Tao Chen, Zhijin Zhao, Xiaoniu Yang · 2023
Signal-to-noise ratio (SNR) is an important indicator for evaluating the quality of the received signals in wireless communication systems and its estimation plays an important role in demodulation and decoding processing. In this paper, we propose an SNR estimation method combining deep learning and covariance matrix (CM). In the method, CM is used as the input of a convolutional nueral network (CNN) to reduce the compuational complexity. We compare our proposed method with other SNR estimate methods, including classical estimation methods and existing deep learning-based approaches. Results show that our proposed method performs the best in low to medium SNR range in terms of estimation accuracy. In high SNR range, it acheives comparable performance with existing deep learning-based estimation methods. More importantly, our proposed method has far lower computational complexity than existing deep learning-based SNR estimation methods.