5G Wireless Channel Estimation Algorithm Based on a New Network Structure

Channel estimation plays an important role in the system performance of wireless networks. With the rapid development and wide application of 5G technology, traditional channel estimation algorithms are difficult to meet the current needs. Deep learning has shown significant improvements in enhancing communication reliability and reducing the computational complexity of 5G. In order to improve the channel estimation, this study models the orthogonal frequency division multiplexing (OFDM) wireless transmission system and proposes a channel estimation algorithm based on REAT-CNN. The algorithm uses an end-to-end multi-layer convolutional neural network (CNN), including residual connection and attention mechanism, to reduce the number of pilots required by traditional channel estimation algorithms. This method not only improves the spectrum utilization, but also improves the reliability of transmission. Through the verification of multiple sets of simulation results, the channel estimation algorithm based on REAT-CNN proposed in this paper can estimate the physical channel more accurately than the least square (LS) method and the linear minimum mean square error method (LMMSE), and reduce the transmission bit error rate (BER) of the OFDM system.