RETRACTED ARTICLE: Deep learning for estimating the channel in orthogonal frequency division multiplexing systems

Abstract

The essential criteria of wireless communication system are accurate signal identification. The channel assessment and adjustment are the two most critical mechanisms for signal identification. Since orthogonal frequency division multiplexing (OFDM) used to decrease bit error rates in current wireless communication systems and to optimize spectral efficiency and compared to a single carrier network, secure channel compensation method. Nevertheless, the system’s quality output is degraded depending on the channel state due to the incorrect channel approximation and noise amplification through the channel compensation procedure. In this paper, a 1D Convolutional Neural Network (CNN) deep learning model to estimate the channel provided and the equalized data also recovered. The Deep Neural Network (DNN) can not only use the channel variation features from previous estimates with correctly chosen outputs, but it can also derive additional features from the pilots and receiving signals. To prove the ability of this model the Bit Error Rate (BER) for the recovered data is compared with the conventional models like Minimum Mean Square Error (MMSE) and Least Square (LS) and compared with the Feed Forward Neural Network (FFNN) model in different digital modulation techniques.