This application belongs to the field of communication
signal processing technology, and more specifically, relates to a method and
system for co-channel interference cancellation based on a multi-star operation neural network. By calculating the equivalent
linear channel response between the target received
signal and the target transmitted
signal, the target linear interference component of the target received signal can be accurately and quickly obtained. This application introduces a target nonlinear reconstruction model with
multiple target single-star models, and then utilizes two feature branches for linear transformation and star operation to quickly and accurately obtain the target nonlinear interference component of the target received signal from multiple dimensions without increasing the stacking of deep convolutional
layers or a large number of recurrent structures. Thus, based on the target linear interference component, the target nonlinear interference component, and the target received signal, effective cancellation of co-channel interference signals in a high-speed dynamically changing channel environment can be achieved.