Strong correlation self-interference cancellation method based on deep neural network

Abstract

The invention provides a strong correlation self-interference cancellation method based on a deep neural network, which introduces a DNN neural network into a strong correlation self-interference cancellation system of a receiving and transmitting system. The method adopts the DNN neural network to fit a self-adaptive cancellation system model, which replaces a traditional self-adaptive filtering algorithm, and puts forward the strong correlation self-interference cancellation method based on the DNN. According to the method, instead of a traditional adaptive algorithm, the system model can effectively eliminate the strong correlation self-interference signals through the trained DNN network model, so that target signals can be recovered from the strong correlation self-interference signals more accurately.

Description

technical field [0001] The invention belongs to the field of strong correlation self-interference cancellation in a simultaneous transmitting and receiving system, in particular to a method for canceling strong correlation self-interference based on a deep neural network. Background technique [0002] The problem of strong correlation self-interference cancellation in simultaneous transmitting and receiving systems has always been a key technology and hot topic. In recent years, adaptive algorithms have been widely used to eliminate self-coupling interference. With the increasing complexity of the electromagnetic environment, it is increasingly difficult to estimate the strong correlation self-interference signal. It is difficult for the traditional algorithm to adapt to the time delay effect and variability of the strong correlation self-interference signal in the receiving signal of the transmitting and receiving system at the same time, and to effectively eliminate the str...

AI Technical Summary

Problems solved by technology

At present, some adaptive filtering algorithms are mainly used to eliminate strongly correlated self-interference signals. With the adaptive change of adaptive filter weights, the error signal will be closer and closer to the target value, but the method of adaptive filtering cancellation will There is a convergence time, and only after the error signal converges can there be a better cancellation result. At the same time, after the superposition of the strongly correlated interference signal and the target signal, the adaptive algorithm loses its effectiveness

In recent years, based on baseband signals and feedback signals, some scholars have used neural networks to reconstruct self-interference signals, and then canceled them with adaptive filters. This method has achieved the elimination of nonlinear self-interference signals, but still requires the use of traditional adaptive filter

Images