Linear neural network based blind equalization

This letter considers the problem of blind equalization in digital communications by using linear neural network. Unlike most adaptive blind equalization methods which are based on matrix decomposition or the Hankel property of matrix, we give a stochastic approximate learning algorithm for the neural network according to the property of the correlation matrices of the transmitted symbols. The network outputs provide an estimation of the source symbols, while the weight matrix of network estimates the inverse of the channel matrix. Simulation results demonstrate the performance and validity of the proposed approach for blind equalization.