OFDM Detection In Double Selectiye Fading Channel And Parallel And Parallel Decoding Algorithm Of Turbo Codes

Multiple input multiple output(MIMO) and orthogonal frequency-divisionmultiplexing(OFDM) were considered as a key technique which will conquer thefrequency selected fading, increase the data transmission rate and improve the systemcapacity in the fourth generation mobile communication. Alamouti space-time codedOFDM system is one of the schemes which combine MIMO with OFDM. As arelatively mature channel coding scheme,turbo coding has aroused wide concernbecause of its excellent performance.In Alamouti space-time coded OFDM system, the simple Alamouti decoding atthe receiver relies on the assumption that the channel do not change over an Alamouticode word period.Unfortunately, when Alamouti-OFDM system is implemented infast and selective fading channel,the Alamouti space-time code orthogonality isviolated as well as the OFDM subcarrier orthogonality,which causes severeperformance degradation.In this paper,with the aid of the new ordering algorithmdevised for sequential interference cancellation, an efficient minimum mean squarederror based staged sequential interference cancellation detector (Staged SIC-MMSE) isproposed to obtain both transmit diversity and Doppler diversity gains. The powerefficiency of the newly proposed detector is comparable with the MMSE basedconventional one-by-one SIC detector, while the computational complexity andprocess delay are greatly reduced. The new ordering algorithm for SIC takesadvantages of not only the instantaneous channel realization as the traditional orderingalgorithm based on post-detection signal-to-interference and noise (SINR), but also theinstantaneous noise realization and the modulation constellation adopted, hence issuperior to the traditional ordering algorithm.The maximum a posteriori probability(MAP) decoding of turbo code needs forrecursive value computation and iteration,for which reason high delay caused limitsthe application in high speed data transmission. Academics have proposed manyparallel decoding schemes to reduce the decoding delay, such as the parallel decodingalgorithm based on predictive control and Sliding window parallel decoding algorithm.With the advantage of low hardware design complexity and the simple controlstructure,sliding window parallel decoding algorithm has received a lot ofattention.Parallel Turbo decoding improves the data throughput, but also reduce the decoding performance. To solve this problem, based on sliding window paralleldecoding algorithm,an improved algorithm is proposed in this paper. Theoreticalanalysis and simulations show that compared with the traditional this algorithm can getbetter performance when having the same parallelism. With the increase ofsignal-to-noise ratio, the advantage of the improved parallel algorithm is moreobvious.Besides, the improved algorithm shows a better convergence than thetraditional when the number of iterations increases.