Research On The Downlink Receiver Technology Of MIMO-OFDM

The target of next wireless mobile communication is to obtain more efficiency ofspectrum. In the wireless communication with the limitation of bandwidth and power,only the implementation of MIMO antennas system can achieve the high efficiency ofspectrum, especially the Blast space-time code with multiplexing gain. At the sametime OFDM technology can reduce the complexity of wide-band receiver detectionand channel estimation. Therefore, the research of this paper is aimed at the downlinkreceiver algorithm design based on the MIMO-OFDM system, and the whole workincludes three parts as below:Research of MIMO detection: In order to obtain the reliable detection underlower SNR, we exploit the iterative detection in MIMO system referred to themulti-user iterative detection. We propose that the structure of iterative detection isdecided by the detection algorithm and channel decoding. Furthermore we utilize theextrinsic Gaussian approximation to conclude that the shorter restriction length ofTurbo code can get better iterative convergence performance than the long restrictionlength. For iterative SoIC (Soft interference cancellation), a hybrid iterative detectiondecoding structure is presented in this paper, and an adaptive MMSE operator basedon the reliability of apriori information is designed to achieve the better advantagethan MRC (Maximum ratio combination) and MMSE. The simulation result of theimproved iterative SoIC shows the performance very close to the optimal MAPiterative detection. In the last part of this section, we extend the iterative detectionto the common multi-input multi-output system with self-interference, and amatch-filter ICI iterative cancellation is proposed in OFDM system, from which wecan see this iterative detection with lower complexity can achieve better performancethan matrix-inverse equalizer.MIMO-OFDM parametric pilots channel estimation: We propose the principle ofpilots design in MIMO-OFDM channel estimation. In order to obtain more precisechannel estimation, we exploit the parametric channel estimation based on thecharacteristic of multi-path channel. Two multi-path allocation algorithms arepresented respectively to obtain the multi-path time delay parameters. One is themulti-path searching algorithm referred to the RAKE receiver in CDMA system, andthe other is ESPRIT multi-path allocation algorithm with MIMO space diversity. Themulti-path allocation can be estimated accurately during a short time. Based on themulti-path allocation estimation, we conclude that the MSE (Mean square error)performance of parametric channel estimation is much outstanding than FFTinterpolation channel estimation. At last an algorithm with verification-symbol isproposed to erase the ambiguity of ESPRIT.Research of MIMO multi-cell multi-user: In this section, we analyze thedifferent channel capacity between MIMO single-user, MIMO uplink MAC(Multi-Access Channel) and MIMO downlink BC (Broadcast Channel). In MIMOBC we propose a partial ordering ZF-DP (Zero forcing Dirty Paper) based on thecorrelation of variant channel to deal with the tradeoff between the multi-userdiversity and equity. In the scenario of multi-cell MIMO downlink channel, a partialZF (Zero forcing) adaptive MMSE filter algorithm is exploited to suppress theinterference of adjacent cells which can achieve lower BER floor than linear MMSEoperator.