On The Channel Estimation And Realization Techniques For LTE Downlink System

As an important 3G Evolution technique, the existing 3G air interface technology has been improved and enhanced by 3GPP LTE Long Term Evolution project. The peak transmission rate can achieve 100Mbps at downlink and 50Mbps at uplink for the 20MHz spectrum bandwidth, In order to support LTE high speed broadband wireless transmission, channel estimation, time and frequency synchronization scheme, as well as the PAPR(Peak to Average Power Ratio) reduction become the critical enabling technologies for LTE communication system. In accordance with the current channel estimation research progress and the LTE downlink protocols analysis, this thesis focuses on the LTE downlink channel estimation and SNR technique, together with their DSP implementation.According to the LTE technical specifications of TS36.211-TS36.214, three channel estimation methods(Least Square(LS), Minimum Mean Squared Error(MMSE) and Best Linear Unbiased Estimation (BLUE)) were analyzed and compared. Then a novel channel estimate method, which based on LS estimate and an improved DFT filter interpolation in frequency domain and linear interpolation in time domain, was proposed for the LTE downlink channel estimation. It is shown that, by configuring the noise threshold properly with the estimated channel information and selecting the filter window according to the distribution of noise and signal energy, the estimation error caused by noise could be reduced. It is validated through simulations that the proposed interpolation based LS channel estimate exhibit reasonable estimation performance with low implementation complexity.As for the SNR estimation problem, after a survey of the Boumard SNR estimation and the MMSE-based SNR estimation algorithms, an improved Boumard SNR estimation algorithm was proposed in this thesis. It is shown through simulations that the proposed improved Boumard offers a reasonable SNR estimation scheme for the LTE system.Due to the demand for practical applications, linear interpolation based LS channel estimation and the Boumard SNR estimation based on AWGN channel are implemented on the DSP platform. The relevant resource usage and estimation accuracy for both algorithms are validated. And our preliminary analysis confirms us that both algorithms are able to achieve reasonable performance with low complexity as well as low system resources occupancy characteristics. Therefore, both algorithms offer a feasible alternative for channel estimation and SNR estimation in the LTE communication system.