The Research Of Channel Estimation Algorithm In LTE Uplink System Based On DFT Transform Domain

LTE Long Term Evolution project is widely known as3.9G technology to enhance thefull range of performance of wireless access technology. It can realize a higher data transferrate, a smaller system delay, and it makes a comprehensive research for the multidimensionalinformation of time domain, frequency domain and airspace. A coherent detection has beentaken in LTE uplink system to demodulate and recovery the transmit signal. The accuracy ofchannel parameters determines the achieve results for coherent detection. It leads to the higherrequirements of channel estimation technology for tracking and estimating the channelinformation. For example, the multipath parameters and the path delay. Therefore, channelestimation technology is still the core technology in LTE uplink system.In this paper, first, it conducts a brief description for the transmission characteristics ofradio channel and the basic concepts about LTE uplink system. By making a comparison withthe structural of OFDM system, it proves that the high PAPR of OFDM system has beenovercome by the application for the structure of SC-FDMA system in LTE uplink. Then itdiscusses classic channel estimation algorithms based on massive pilot distribution of LTEuplink and conducts a more intuitive analysis of their respective characteristics throughsimulation. And then in order to elicit the focus of this paper, it summarizes the originalimproved programs of LS algorithm which is easy-to-use. The focus content is the improvedchannel estimation algorithm based on DFT transform domain in LTE uplink system.Channel estimation algorithm based on DFT transform domain is also a kind of the noisereduction program for LS algorithm. It takes full advantage of characteristics for the signalenergy concentrated in time domain to make a noise separation. In this paper, channelestimation algorithm based on DFT transform domain in LTE uplink system has beenconducted a comprehensive analysis. It summarizes the original algorithm based on DFTtransform domain. The original algorithm can be divided into two programs based on thedifferent processing angle of the different improved programs. One is the noise reductionprogram by setting the threshold and the other one is processing by the window function, theycan be combined with each other. It writes from another angle that making a research to thetruncation length of the low-pass filtering in time domain. And it proposes an improvedprogram with the adaptive truncation length of the low-pass filtering. According to thechanges of the signal energy, the improved algorithm proposed in the paper adaptively judgesand obtains the more accurate length of the signal energy concentrated in time domain. Itmeets the requirements of the actual application system. Thus, it takes a more accurateseparation of effective signals and noises in the estimation process, solves the estimation errorof the channel mismatch, and improves the accuracy of channel estimation result. Finally, with the adaptive idea proposed in this paper, it solves the pre-estimation problem forautocorrelation matrix of LMMSE algorithm, and proposes a simple and practical adaptiveprogram. To obtain the autocorrelation matrix adaptively, the improved program is no longerbased on a large number of statistical information and the pre-estimation for autocorrelationmatrix is more accurate. In general, the improved algorithm reduces the complexity ofLMMSE algorithm to some extent.