Wireless communication method and system based on deep learning under nonlinear distortion condition

Abstract

The invention discloses a wireless communication method and system based on deep learning under a nonlinear distortion condition, and the method comprises the steps: carrying out the channel estimation of an equivalent channel of a received signal through an LS algorithm or an MMSE algorithm, so as to obtain an estimated signal; demodulating the estimated signal through a deep convolutional neuralnetwork to obtain a demodulated signal; and training the demodulated signal through a deep learning model, and outputting an estimated value of the received signal. According to the method, a receiver architecture based on deep learning is adopted, and an initial channel estimation matrix is corrected through a large number of training data sets; and meanwhile, a deep learning network model is used for matching nonlinear distortion factors, so that the safety capacity of the system is improved.

Description

technical field [0001] The present invention relates to the technical field of wireless communication, in particular to a wireless communication method and system based on deep learning under nonlinear distortion conditions. Background technique [0002] Due to the extensive use of components such as antennas, power amplifiers, and mixers in the wireless communication system, high-order components will be introduced into the transmitted wireless signal, resulting in nonlinear distortion of the signal, which will lead to a decrease in the link quality of the wireless communication system. cause system capacity reduction. At present, for the problem of nonlinear distortion, digital predistortion (DPD) processing is generally used in the transmitter to compensate the signal distortion of radio frequency devices, but digital predistortion technology often has the disadvantages of difficult modeling, complex calculation and poor real-time performance. As a result, its applicatio...

AI Technical Summary

Problems solved by technology

The present invention constructs a network model based on deep learning by using a large number of training data sets, and only needs to train the demodulation module at the receiving end, so as to obtain a deep learning network model that matches the nonlinear distortion, and improves the wireless communication system capacity, which overcomes the shortcomings of difficult modeling, complex calculation and poor real-time performance in the prior art

Images