Research On Key Techniques In Wideband Mobile Communication Systems

With the rapid growth of wireless multimedia services and Internet services, the past few years have seen the transition of wireless communications from narrow band and low data rate to wide band and high data rate. Block transmission including orthogonal frequency division multiplexing (OFDM) and single carrier with cyclic prefix (SCCP) is one of the key transmission techniques in wideband wireless communications. In mobile applications, however, such transmission technique is very sensitive to channel time variation, as the channel time variation will destroy the orthogonality among sub-channels and therefore degrade the system performance. In order to cope with the channel time variation, this paper makes tries of finding solutions to this problem and the following are the results:1. In pilot-aided time-varying channel estimation for OFDM systems, pilot can be multiplexed either in the time domain or in the frequency domain, and both the cases have been considered in this paper. In view of a pilot-aided channel estimation scheme proposed recently, which is based on the estimation of symbol-averaged channel impulse response (CIR), we first analyze its defects in parameter selection and system performance, and then, relying on a general basis expansion model of time-varying channels, we make improvements to it. The improved scheme features enhanced performance and a more general form. For the latter case, relying on the idea of the pilot symbol aided modulation in conventional single carrier system and the Karhunen-Loeve (KL) expansion of time-varying channels, we first present a channel estimation scheme for OFDM and then show theoretically that it has the same performance as the Wiener interpolation based channel estimation.2. An enhanced EM-based iterative channel estimator for coping with channel time variation is proposed for mobile multiple-input multi-output orthogonal OFDM (MIMO-OFDM) systems. In this scheme, by using data decisions obtained in iterations, the ICI are constructed and then cancelled from the received signals in order to reduce their impactions on channel estimation. In addition, an improved predictor is proposed to refine the channel initial estimates.3. Channel estimation for subband orthogonal frequency division multiple access (OFDMA) operating over time-frequency selective channels has been considered. By defining a so called BEM in the frequency domain, two signal models that are fitted into the framework of the EM algorithm are derived for subband OFDMA systems operating respectively in the signal and the multiple transmit antennas cases, and as a result, two EM-type channel tracking schemes are proposed for the two applications, respectively. As compared with the conventional linear interpolation scheme, the proposed schemes are more robust to the channel time-frequency selectivity. Moreover, the subband size of our proposed schemes can be flexibly adjusted according to the throughput requirement of a specified user.4. The channel estimation and equalization for SCCP over doubly selective channels have been studied. The main contributions are to propose a doubly iterative receiver that consists of a minimum-mean-squared-error turbo equalizer with soft interference cancellation and an EM-based iterative channel estimator. The joint detection in the time domain makes the proposed equalizer different from existing turbo equalizers for SCCP. For this reason, the finite support of channel length can be used to reduce the computational complexity, and similarly, the joint detection can avoid the limitation of separate detection caused by ill-conditioned matrix. In order to obtain reliable CSI that is indispensable in the channel equalization, we use the data decisions over multiple blocks to estimate the CSI iteratively relying on the EM algorithm. The proposed estimation scheme makes a clever use of CP symbols to initialize the EM algorithm, which therefore results in improved estimation performance and reduced complexity as well.5. For the purpose of mitigating ICI caused by channel time variation in OFDM systems, a doubly iterative equalization scheme that consists of two serial equalizers is proposed. By exploiting the approximately banded structure of channel frequency response (CFR) matrix, a turbo equalizer in the first stage is used to cancel the ICI in the out of the band, and the other turbo equalizer in the second stage aims to suppress the ICI inside the band. By these two steps, the ICI can be effectively mitigated and the complexity can also be reduced. Moreover, the scalability of band size of both the inside and the out of the band enables a good tradeoff between performance and complexity. Theoretic analysis indicates that the proposed technique takes a general form, and many equalization algorithms can be seen as a special case of our algorithm.