Dissertation > Industrial Technology > Radio electronics, telecommunications technology > Communicate > Data communication > Data transmission technology

A PSAM Channel Estimation Based on DFT and DCT for OFDM System

Author JingLei
Tutor ZhaoXiaoZuo
School Jilin University
Course Signal and Information Processing
Keywords Orthogonal frequency division multiplexing ( OFDM ) Channel estimation LS LMMSE SVD DFT DCT
CLC TN919.3
Type Master's thesis
Year 2011
Downloads 183
Quotes 1
Download Dissertation

OFDM technology with anti-interference ability, high spectrum efficiency, using IDFT / DFT transform to achieve the modulation and demodulation, etc., is an effective high-speed data streaming mode. OFDM technology Because of these unique advantages, it has been used as a standard is widely used in outdoor wireless LAN, among the European Digital Video Broadcast and Digital Audio Broadcasting. The shortcomings of the OFDM system is very sensitive to frequency deviation. In poor wireless mobile communication propagation environment, the transmission symbol will occur amplitude and phase distortion, which limits the transmission speed of the data symbols. Reliable channel estimation is one of the key OFDM system to achieve high-speed data transmission, it is the basis of dynamic bit allocation and signal coherent detection. To improve the performance of the OFDM system, we must channel estimation techniques to obtain accurate channel state information. OFDM channel estimation methods can be divided into three categories: the first category is the The pilot assisted (PSAM) channel estimation, it is inserted in the data symbols the pilot to estimate the channel frequency response (CFR). Although the algorithm is easy to achieve, but it takes up a lot of spectrum resources, but it also brings the inevitable delay; second category is based on the decision feedback (DD) channel estimation, a or symbols The channel response estimation value estimated for the current time and the channel response. In order to improve the performance of channel estimation, the algorithm will usually make use of pilot symbols; blind channel estimation (BCE), only the statistical characteristics of the transmission symbol to estimate the channel. The method does not require insertion of pilot symbols, save bandwidth, but the slow rate of convergence and the computation is also more complicated. This paper focuses on PSAM channel estimation algorithm and its performance, and for the simulation analysis in MATLAB. Paper guide for different frequency insertion pattern, first discussed the massive pilot, pilot comb and rectangular Pilot insertion; followed by a detailed analysis of LS, LMMSE, SVD, DFT and DCT-based channel estimation algorithm; constant value interpolation, linear interpolation, Gaussian interpolation Cubic interpolation and DFT interpolation algorithm to do the introduction; Finally, through computer simulation, compared and analyzed under different pilot patterns, application of different channels estimated channel algorithm and the interpolation to obtain the frequency domain response. Simulation results showed that: the best the LMMSE algorithm of estimation performance, but the complexity of the algorithm is also the largest, is difficult to achieve; due to the matrix dimension reduction algorithm based on SVD calculation is greatly simplified, the performance is almost the same LMMSE; LS algorithm performance a bit weak, but the calculation is simple and easy to implement; DFT and DCT algorithm-based BER \In this paper, the comb pilot \Estimated based on the IDFT / DFT channel algorithm, first LMMSE criteria pilot position of the channel frequency domain response as the IDFT transform the input then the CFR will be obtained, and then the thus obtained channel impulse response in the time domain linear transformation, Finally, calculated by DFT interpolation all sub-carriers of the CFR. The algorithm still exists the BER curve \Therefore, further improvement to switch to the linear interpolation method to obtain CFR, i.e. the channel impulse response (CIR) of the first pilot positions transformed back to frequency domain after DFT operation, and then in the frequency domain linear interpolation calculation, thereby obtaining the BER curve completely eliminates the \DFT / IDFT-based channel estimation algorithm is the the LMMSE standard pilot CFR after the DFT operation, and the resulting data sequence in the transform domain low-pass filter and zeros, and then back to the frequency domain by IDFT interpolation transform . LMMSE criteria of the algorithm, so the calculation than the larger, but did not significantly improve the estimation performance. This is caused by the \IDCT / DCT-based channel estimation algorithm by LMMSE norms CFR first IDCT operation multiplied by the gain factor and then into the time domain, and then in the time domain CIR trailing zeros, then the DCT interpolation. IDCT and DCT transform of the different points, so the last in the frequency domain, and then multiplied by a gain factor prior to compensate for the mismatch. Nevertheless, the estimated effect is still not ideal. Estimate the channel based on DCT / IDCT algorithm the LMMSE criterion pilot CFR obtained after conventional DCT operation, the end of the corresponding spectral sequence in the transform domain, any zero-padding, and then through the expansion transform the IDCT (Eidct) Back to the frequency domain to obtain all the subcarriers of the CFR. The algorithm for carrying out a DCT transform, resulting in data shift occurs, and therefore can not be used directly IDCT transform should be used EIDCT be compensated. The resulting channel estimation performance greatly improved, reducing the system BER. We simulated and compared the improved channel estimation algorithm performance under different system and channel parameters. Simulation results show that: the small amplitude phase modulation method to increase the number of pilot to select a larger number of subcarriers of the FFT transform to reduce the Doppler frequency shift, to reduce the path number, choose an integer sampling channel, and a more accurate selection interpolation method will improve the accuracy of channel estimation, while improving system performance.

Related Dissertations
More Dissertations