Fading Channel Prediction Based on Self-optimizing Neural Networks

Channel prediction is an important technique for compensating fading channel in mobile communications. We proposed a channel prediction method based on complex-valued neural networks as a previous work. In this paper, we introduce a penalty function to the weight update in the complex-valued neural network to realize a learning dynamics that can self-optimize network structures according to fast changing communication environments. This presents an adaptive and highly accurate channel prediction method. We demonstrate the ability of the proposed method in a series of simulations.