Channel machine learning estimation method based on intelligent iterative initial value selection

Abstract

The invention relates to a channel machine learning estimation algorithm based on intelligent iterative initial value selection. The algorithm mainly comprises the following steps: (1) modeling a probability model of a channel by using a Gaussian mixture model (GMM); (2) performing channel estimation by using optimal Bayesian parameter estimation; (3) determining an initial value of an iterative process by using an improved K-means algorithm; (4) solving the edge probability density function involved in the step (2) by using an approximate message passing algorithm (AMP); and (5) iteratively solving parameters in the Gaussian mixture model by using an expectation maximization (EM) algorithm. According to the invention, the sample information of the channel is fully utilized and clustered,a partial iteration process of an expectation maximization algorithm (EM) is replaced to ensure a faster convergence rate and better MSE performance of iteration, and meanwhile, the channel is estimated by adopting an optimal Bayesian parameter estimation algorithm according to the sparse characteristic of the beam domain channel gain.

Description

technical field [0001] The invention belongs to the field of wireless communication technology, specifically a channel machine learning algorithm based on intelligent iterative initial value selection, based on the sparse characteristics of channel gain in the beam domain, and using machine learning related algorithms to improve Bayes-GMM Channel estimation algorithm to estimate the channel. Background technique [0002] With the development of information technology in today's society, the development of wireless communication is also becoming faster and faster. From the first generation mobile communication system (1G) to the fifth generation mobile communication system (5G) with great development potential, the quality and speed of communication are getting higher and better, but the traditional technology It has gradually failed to keep up with people's demand for mobile communication services, so one of the core technologies of 5G - massive multiple-input multiple-outp...

AI Technical Summary

Problems solved by technology

In the current research on channel estimation using machine learning, there is only one channel estimation algorithm based on Bayes-GMM, which can obtain good mean square error (MSE) performance, but the Gaussian mixture model (GMM ) in the iterative initial value only adopts an average selection method. It is precisely because of this selection method that the convergence speed of the channel estimation algorithm is very slow, and the channel sample information is not fully utilized

Images