Study Of Key Technologies For MIMO Wireless Communications Systems

In order to meet the requirement of high transmission rate, wireless terminals will be equipped with multiple antennas for future wireless communications systems, ting in the so-called multiple input multiple output system, i.e.. MIMO system. With the implement of the MIMO techniques, the system capacity can be significantly improved。Furthermore. MIMO system can provide multiplexing gain, diversity gain and so on. On the other hand, if the downlink transmission from the BS to the MSs is significantly degraded by the shadowing effect, the signal is very weak. In this case, the wireless relay station (RS) can be used to extend the coverage area. The basic principle of wireless RS is to process the signal from the source node and then send it to the destination node. According to different working scheme, the RS can be classified as amplify-and-forward (AF) and decode-and-forward (DF). The AF is also called non-regenerative relay, where the RS only amplifies or processes the received signal linearly and then forward it to the destination node. The DF is also called regenerative relay, where the RS decodes the received signal and then recode the information bits. After that, the RS sends it to the destination node. Compared to the FF scheme, the DF one can obtain higher diversity gain. However. DF scheme imposes higher requirement on processing capability of RS and also may cause trouble on the security. The wireless relay technique can be effectively combined with the MIMO technique based on the development of MIMO technique, yielding the MIMO wireless relay technique, which attracts the intensive research effort. Based on this background, in this thesis. MIMO wireless communications has been intensively studied and the major contributions are as following:1. For a multiple user MIMO (MU-MIMO) uplink transmission system, we proposed the block diagonalization (BD) based parallel and successive MUI mitigation algorithms. For parallel BD algorithm, the BS projects the received signal intended for one user into the null space of other users, so that the received signal intended for one user is orthogonal to other users. In this case, different users" signal can be separated completely and detected independently. For successive BD algorithm, the BS assumes that the detection for the previous users is completely correct. Consequently, the interference imposed by the previous users can be completely removed from the received signal. For both parallel BD and successive BD, the MU-MIMO system is decomposed into several single user MIMO (SU-MIMO) systems. Hence the existing algorithms for SU-MIMO can be used directly. The proposed algorithms can remove the MUI while keeping the joint processing capability provided by the multiple receive antennas, hence the system performance can be effectively improved.2. For MU-MIMO downlink transmission, we proposed the robust successive MMSE (SMMSE) preprocessing technique to mitigate the MUI. For preprocessing technique, the transmitter side can only get the CSI by feedback, which results in the error between the estimated CSI at the transmitter side and the perfect CSI. The error can be modeled as additive white Gaussian noise (AWGN) with zero mean. Considering the statistical property of the error, the robust SMMSE preprocessing technique is designed. The simulation results proved that the roust SMMSE algorithm outperforms the traditional one. and even completer removes the error floor.3. For MIMO non-regenerative wireless relay system, we studied the optimal RS power allocation, which can meet the given MMSE constraint for the MMSE receiver at the MS. We transformed the question into a convex optimization problem and derived the feasible range of MMSE. Furthermore, by using water-filling algorithm, we obtain the optimal power allocation coefficients. The simulation results showed that the required power is reduced while reducing the number of antennas at RS.4. For MIMO non-regenerative wireless relay system, we proposed the joint transceiver design for BS, RS and MS when nonlinear MMSE decision feedback equalizer (DFE) is used at the MS. Firstly, the linear matrix inequality (LMI) is used to get the optimal feedforword and feedback matrix at the MS. Afterwards, the geometric mean decomposition (GMD) of matrix is used to get the optimal linear weighting matrix at RS and the optimal preprocessing matrix at BS. The simulation results showed that the proposed optimal joint non-linear MMSE-DFE transceiver significantly outperforms the traditional optimal joint linear MMSE transceiver.5. For MIMO non-regenerative wireless relay system, we proposed the block diagonalization (BD) based MUI mitigation algorithm for the scenario where the BS communicates with multiple MSs via a RS. Firstly, the RS invokes the BD based receiver to process the received signal from BS, so that the signal for different users is separated completely. Afterwards, the BD based preprocessing technique is adopted so that each user can only receive the signal intended for itself. In this case, the MUI has been completely removed for each user. Finally, one MU-MIMO wireless relay system is decomposed into several SU-MIMO ones. Furthermore, we also studied the optimal power allocation at RS to achieve the maximum attainable capacity.