Channel state information (CSI) plays a vital role in fading-adaptive wireless systems, whereas acquired CSI is generally inaccurate due to feedback delay. Channel prediction that is able to forecast upcoming CSI provides a promising approach to tackle this problem. Exploiting the capability of time-series prediction in neural network, a fading channel predictor is proposed in this paper. Since the number of neurons in the input and output layers is tunable to adapt the number of antennas, the predictor well suits multi- antenna systems. Numerical results in antenna selection system reveal that a substantial performance gain can be achieved by applying channel prediction.
Paper
Full text
Neural Network-Based Channel Prediction and Its Performance in Multi-Antenna Systems
Semantic Scholar · Engineering · 2018
Abstract
Channel state information (CSI) plays a vital role in fading-adaptive wireless systems, whereas acquired CSI is generally inaccurate due to feedback delay. Channel prediction that is able to forecast upcoming CSI provides a promising approach to tackle this problem. Exploiting the capability of time-series prediction in neural network, a fading channel predictor is proposed in this paper. Since the number of neurons in the input and output layers is tunable to adapt the number of antennas, the predictor well suits multi- antenna systems. Numerical results in antenna selection system reveal that a substantial performance gain can be achieved by applying channel prediction.