Research On Channel Quality Measurement And Prediction Algorithm For LTE Uplink

With the development of LTE technology, to improve spectrum efficiency and power efficiency is increasingly urgent. According to the channel conditions, the link adaptation technique can select the best transmission power and the encoding modulation parameters, which can achieve the maximum power efficiency and spectrum efficiency.The main purpose of this paper is to improve the spectral efficiency of LTE uplink system, and the key technologies of adaptive modulation and encoding are discussed and studied. In the adaptive link, BS needs to send feedback information to UE. In order to ensure the real-time and validity of the feedback information, we need measure and predict the channel quality. Channel quality prediction can turn the predicted channel information into effective adaptive indicators. The system can overcame the feedback delay through forecasting the channel equality and achieve a balance between system throughput and block error rate.Firstly, this paper presents the frame structure of time division duplex mode in LTE standard protocol and key technologies of LTE adaptive link, which provides a signal model and theoretical basis for the following research. Channel prediction and channel quality measurement are based on channel estimation, and the channel estimation method used in this paper is MMSE. A theoretical analysis of the traditional multi-carrier channel quality measurement method is given. Compared in the same case, the EESM algorithm and the improved combined algorithm have stronger applicability than other multi-carrier channel quality measurement methods. For the single carrier characteristics of SC-FDMA signal in LTE uplink, a single carrier method is proposed to measure the quality of the single carrier channel with the error vector magnitude(EVM). Based on EVM algorithm, an improved single carrier frequency measurement algorithm is proposed. The improved algorithm can enhance the throughput in the same situation.In the TD-LTE system, there is a certain gap between uplink subframes, so there exists a certain delay from the feedback uplink channel quality measurement information to reach the user side. Therefore, it is necessary to predict channel quality for overcoming the channel variation caused by delay. Channel quality depends on the channel change,interference, noise and other factors. Due to the superposition of these random factors, the channel quality has weak correlation in time, and there is a big challenge for prediction. In the case that other factors are relatively stable, we can achieve channel quality prediction by predicting the channel response. In this paper, a theoretical analysis of the traditional linear and nonlinear forecasting methods is carried out. The innovative application of genetic programming in machine learning field is applied to the channel prediction. Genetic programming algorithm can bring a certain degree of forecasting performance benefits at the price of high computational complexity.