Research On Channel Estimation And Equalization Algorithm For Long Term Evolution Uplink

Orthogonal frequency division multiplexing (OFDM) is well-suited totransmission through mobile radio channels subject to frequency selective fading dueto multipath propagation. Single carrier frequency division multiple access(SC-FDMA), which utilizes single carrier modulation, discrete Fourier transform(DFT) precoded OFDM, and frequency domain equalization, is a technique that hassimilar structure and performance as orthogonal frequency division multiple access(OFDMA). One prominent advantage over OFDMA is that SC-FDMA has better peakpower characteristics because of its inherent single carrier structure. Recently,SC-FDMA has been adopted for the uplink multiple access scheme in third generationpartnership project long term evolution (3GPP-LTE), and greatly benefits the mobileterminal in terms of transmit power efficiency and manufacturing cost.Firstly, this paper studies on LTE physical layer technical specifications andSC-FDMA signal processing. After the fading characteristics of the radio channel areanalysed, we focus on the channel estimation in physical uplink shared channel(PUSCH). Due to the minimum mean square error (MMSE), liner minimum meansquare error (LMMSE) and singular value decomposition (SVD) are generally toocomplex for a practical implementation. In multipath channels withnon-sample-spaced time delays, the conventional DFT approach will cause powerleakage and result in an error floor. A new discrete cosine transform (DCT)-basedchannel estimation algorithm using wavelet denoising technology for LTE uplinksystems is proposed. Mallat algorithm, which can achieve binary discrete wavelettransform (DWT) coefficients conveniently, is employed for channel frequencyresponse accuracy improvement before DCT. DCT-based channel estimator has thecapacity to cope with aliasing error caused by DFT-based method. Then computersimulations verify that the proposed algorithm can achieve bit error rate (BER)benefits and reduce mean square error (MSE) performance as compared with theconventional estimators.Equalization is an essential compensation for channel distortion caused bychannel frequency selectivity. Frequency domain linear equalization suffers fromsignificant noise enhancement for transmission over deep frequency-selective fadingchannels. An iterative block decision feedback equalization (IBDFE) based on the MMSE criterion outperforms existing DFEs. However, in each iteration of thefeedback, the cross-correlation function between the detected symbols and thetransmitted symbols has to be calculated. Hence, the system complexity of the IBDFEis quite high. In this paper, a joint channel estimation and simplified iterativefrequency domain DFE is proposed. DCT-based channel estimation algorithm is usedto achieve the channel frequency respond (CFR). According to the channel stateinformation, a reasonable decision error probability can be predefined. As a result, incomparison with the IBDFE, our proposed scheme significantly reduces thecomputational complexity required.