Channel Quality Prediction In Link Adaptation

Although the frequency resources are infinite,the available frequency bands are limited.At present,as most of the available frequency bands have been allocated,it is a key problem in efficiently using the spectrum resources.Link adaptation technology can effectively improve spectrum utilization in mobile communication and wireless communication.Because single carrier-frequency division multiple access(SC-FDMA)technology can effectively reduce the peak-to-average ratio of communication signals compared with orthogonal frequency division multiplexing(OFDM),in the fourth generation and the fifth generation,many experts have carried out a lot of research on it.This thesis gives a detailed description of the link adaptation implementation of the SC-FDMA communication system,including obtaining channel information from the SC-FDMA frame,mapping the channel information into effective signal-to-noise ratio,and selecting the modulation and coding scheme(MCS)according to the predicted effective signal-to-noise ratio.Considering that the channel condition of the scattering conmunication is very bad,this thesis chooses to study the link adaptive technology in the scattering communication system.The channel quality prediction is an essential step in link adaptation and it is studied chiefly in this thesis.There are two key steps in channel quality prediction which are extracting specific components in time series and selecting appropriate prediction algorithms for time series.And this thesis introduces data decomposition algorithms and prediction algorithms in detail.The variational mode decomposition(VMD)algorithm is selected to decompose the data and the back propagation(BP)neural network is selected to predict the channel quality.In addition,a multi-step prediction error estimation method for the BP neural network is proposed,which provides a reference for the multi-step prediction performance of channel quality.In machine learning,ensemble learning can be used to combine base learners,which is more effective than using base learner alone.This thesis proposes a channel quality prediction method which is based on VMD and Bagging.Compared with other existing methods,this prediction method can reduce channel error and improve accuracy for the channel quality prediction.In link adaptation based on SC-FDMA system,the throughput of the system can be improved and spectrum resources can be further utilized.