Channel estimation method based on improved GAN network in large-scale MIMO

Abstract

The invention relates to a channel estimation algorithm based on an improved GAN (Generative Adversarial Network) in a large-scale MIMO (Multiple Input Multiple Output) system, and provides a channel estimation method adopting an improved GAN (Generative Adversarial Network) in order to improve the performance when channel estimation based on deep learning is carried out in a one-bit uplink multi-user large-scale MIMO system. In the method, a random quantization method is introduced to improve the input of the GAN network, so that the input data is more real; penalty terms are respectively introduced into a generator and a discriminator to generate a new optimization objective function, so that the network optimization direction is correct, and a network structure is determined through model simulation. The GAN network learns non-trivial mapping from quantitative measurement values to channels by using priori channel estimation observation values; and adversarial training is carried out on the generator and the discriminator to predict a more real channel. A numerical simulation result shows that the method obviously improves the channel estimation accuracy of the large-scale multi-input and multi-output system from the angle of a normalized mean square error (NMSE), and the channel estimation accuracy of the large-scale multi-input and multi-output system is improved by the aid of the method in the aspect of the NMSE (Normalized Mean Squared Error) of the large-scale multi-input and multi-output system.

Description

technical field [0001] The invention relates to the technical field of wireless communication, in particular to a channel estimation method based on an improved GAN network in massive MIMO. Background technique [0002] In future wireless communication systems, massive multiple-input multiple-output technology is one of the key technologies to improve system capacity and spectrum utilization. By deploying a large number of antennas in the base station, the massive MIMO system not only improves the multiplexing capability of spectrum resources among multiple users, but also greatly improves the data transmission rate due to its strong anti-interference ability; however, the current massive MIMO system is usually equipped with high Resolution analog-to-digital converter (ADC, Analog to Digital Converter), which results in high power consumption and hardware complexity. To solve this problem, existing technologies use massive MIMO with one ADC as an alternative solution. The ...

AI Technical Summary

Problems solved by technology

However, one of the reasons GANs are difficult to train is that they are prone to mode collapse, that is, they only learn features that describe a few distribution modes

Images