Channel Estimation for LTE downlink

Abstract

In LTE system, supporting high mobile user speed is one of the key of requirements. However, the channel variation in different mobility scenarios is a significant challenge to achieving this goal. The channel estimation is required at the receiver part to satisfy the good performance with channel variation. In this master thesis, I use the different estimators to estimate the channel at the LTE Downlink system with different scenarios. The LS block fading estimator use at low mobility where the channel is statistic during one subframe. When the user speed more than 20 km/h the LS block estimation degrade in MSE and throughput performance. The LS fast fading outperforms the LS block fading at the velocity more than 25 km/h. LS estimation can be simply implemented with low computational complexity, however the MSE performance is not satisfactory at the high mobility. The LMMSE method is optimum in minimizing the MSE of the channel estimates and in throughput performance.However, the algorithm contains matrix inverse operation and other complex operations which causes high computational complexity. To compatible between these different estimators with good performance and low complexity, the proposed method is used. This method combines three different type channel estimations which is termed Hybrid Linear Mean Square Error ( HLMSE). This proposed estimator is a hybrid of the Least Square (LS) estimator block for low mobility , LS fast fading estimator for moderate mobility and Linear Minimum Mean Square Error (LMMSE) estimator for high mobility. The performance of the HLMSE estimator compared with LS in terms of throughput and Mean Square Error(MSE) outperforms the LS in both throughput and MSE. The complexity of the HLMSE can be controlled by channel variation, which depends on mobility.