Channel Prediction Algorithms For Time-varying MIMO Systems

By using multiple antennas at both transmitter and receiver, MIMO(Multiple-Input Multiple-Output) technique has the ability to improve system spectral efficiency without increasing bandwidth or transmit power. The transmitter needs the accurate channel state information for the upcoming transmission frame to use pretreatment technology, such as water powering, beamforming, to achieve better system performance, and then performing adaptive transmission methods at the moment according to the channel state information. However, CSI fed back to the transmitter becomes outdated due to the channel variation. This thesis addresses channel prediction for compensating in MIMO systems. As follows are the main contributions:1. The complex BP neural network based prediction is derived. The algorithm is good at dealing with strong nonlinear characteristics of the data. Simulation results show that the proposed method can effectively compensate the channel performance compare with LS-SVM algorithm. And it has a better prediction effect.2. A new channel prediction method is proposed which is based on complex wavelet neural network in MIMO systems. The algorithm is a combination of wavelet theory and complex neural networks.Simulation results show that the proposed method has better noise tolerance and prediction effect compared with traditional prediction algorithms. such as complex BP neural network algorithm and LS-SVM algorithm.3. The optimizing of the step size is proposed in Grassmannian Predictive Coding algorithm. And then gives GPC algorithm based on Lloyd quantization. At last, The prediction of the channel state information in the next time based on the index. Simulation results show that the proposed method the method has a better prediction effect.