Prediction channel modeling method based on adversarial network and long short-term memory network

Abstract

The invention discloses a predictive channel modeling method based on a generative adversarial network and a long short-term memory artificial neural network, which effectively realizes a channel prediction function under different frequency bands and scenes and generates a large number of channel data sets for a simulation experiment. The method comprises the following steps: firstly, inputting channel measurement data of existing frequency bands and scenes for training; secondly, learning real channel data by using a long-short-term memory artificial neural network to obtain channel time sequence features; through adversarial learning of the generative adversarial network, redundant information of channel data is greatly eliminated, accurate channel data is generated according to measurement data, and massive channel information is obtained. And finally, obtaining the balance between the generative model and the discrimination model in continuous iteration of the generative adversarial network, and outputting a trained predictive channel model. The channel statistical characteristics obtained by model prediction can clearly illustrate the prediction learning of the method on the channel distribution characteristics, and the real-time and complex prediction problem in wireless communication can be solved.

Description

technical field [0001] The invention belongs to the technical field of channel modeling, and in particular relates to a prediction channel modeling method based on an adversarial network and a long short-term memory network. Background technique [0002] With the development of new technologies and applications in the sixth generation wireless communication system (The sixth generation, 6G), traditional passive channel characterization brings some problems, such as high channel measurement cost, complex channel parameter estimation, and unpopular channel models. The complex and diverse scenarios in the 6G standard will require high-performance channel detectors for measurement, but these instruments are very expensive, and channel measurement cannot exhaust all frequency bands and scenarios. When channel parameter estimation is performed, the amount of channel data to be processed is very large, and the algorithm complexity is very high. Finally, traditional non-predictive ...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to provide a predictive channel modeling method based on generative confrontation network and long short-term memory artificial neural network to solve the problem of insufficient data sets in channel modeling, low quality and diversity of required channel data, and parameter generation efficiency low technical issues

Images