Research On Transceiver Design For Multi-user MIMO Relay Systems

It is well-known that multiple-input multiple-output (MIMO) technique can greatly improve the spectral efficiency and link reliability without any additional bandwidth and transmit power, by exploiting multiple antennas at both transmitter and receiver. It can be seen as a critical technique to meet the increasing demand of high-quality high-rate wireless services. On the other hand, recent research results have showed that relaying technique can offer significant benefits, including throughput enhancement, power re-duction and coverage extension. It is therefore considered as a promising candidate for various wireless standards, such as LTE-Advanced and IEEE802.16m. Thus, the combi-nation of MIMO and relay can be regarded as a very important component for the future wireless communication systems. The performances of MIMO relay systems can be fur-ther improved by employing the proper transceiver design with the known channel state information (CSI). Therefore, this thesis studies the transceiver design for the downlink multi-user MIMO relay in the context of the next generation cellular systems. In order to dig the performance gain brought by MIMO and relaying techniques, the detailed design schemes and the performance analysis are given under the perfect and imperfect CSI, respectively. The main contributions of the thesis are summarized as follows:1) The transceiver design with perfect CSI is studied based on different criteria. Firstly, for sum mean-square-error (MSE) minimization, the analytical solutions of Lagrange multipliers and transceiver matrices are derived exactly by modifying the transceiv-er structure. Then, an iterative algorithm inspired by the alternating optimization is developed to realize the transceiver design. It is shown that the numerical search for Lagrange multipliers within the traditional structure is eliminated. Compared to the existing works, the proposed scheme yields almost the same performances, but with a lower complexity. Secondly, for weighted sum rate maximization, the original prob-lem is converted into an equivalent weighted sum MSE minimization problem. Then, it is further decoupled into different convex sub-problems, and an iterative algorithm is presented to obtain the local optimal solution. It is shown that the proposed scheme have much better performances than the existing works. Thirdly, assuming that the maximum transmit power of source and relay is employed, the associated problem-s are transformed into unconstrained ones. A relationship between the gradients of weighted sum rate and weighted sum MSE is established by analyzing their differ-entials. Then, the gradient expressions of transceiver matrices are derived, and the design is achieved by applying the gradient descent algorithms.2) The robust transceiver design with stochastic CSI errors is addressed. Two robust design schemes for different applications are proposed to minimize the sum MSE. The first scheme decouples the original problem into different sub-problems, where the convexity of each one is verified, and the optimal solution is solved. Afterwards, an iterative algorithm is proposed to obtain the local optimal solution. In the second scheme, it is proved that the sum MSE can be expressed as the sum of two values, each of which stands for the MSE in the first and second phase, respectively. Based on this fact, the original problem is decomposed into two sub-problems, and each one is further relaxed to a scalar-variable power allocation problem by using the majorization theory. The approximate closed-form solution is obtained by utilizing the water-filling algorithm. Compared to the first scheme, it is shown that the second scheme has a little bit degraded performance but a greatly reduction in both the computational complexity and signaling overhead. Thus, it can be concluded that the first scheme is more suitable for the low-speed or fixed scenarios, while the second scheme can be applied to the high-speed scenario.3) The uplink-downlink duality result is extended from the perfect CSI case to the imper-fect CSI case. By employing the KKT conditions, it is demonstrated that sum MSE in the downlink system is identical to those in the uplink system, and vice versa, which can be achieved under some certain conditions. Compared to the existing methods, this one is simpler. Moreover, under the independent transmit power constraints and the total transmit power constraint, the transformations between the uplink and down-link transceiver are derived, respectively. In order to validate the obtained duality result, an uplink robust transceiver design scheme is proposed.4) The worst-case robust transceiver design with bounded CSI errors is investigated. Two robust design schemes are proposed to minimize the sum MSE and the total transmit power, respectively. In the first scheme, the infinite constraints are transformed into the finite linear matrix inequality constraints using the generalized Sign-definiteness lemma, and then the original problems are converted into different semidefinite pro-gramming (SDP) sub-problems with respect to each variable. An alternating iterative algorithm is developed to obtain the local optimal solution. In the second scheme, according to the idea of cutting plane, the associated problems are solved utilizing an iterative procedure which consists of alternating transceiver design and channel deter-mination steps. Moreover, several extensions of the proposed schemes are discussed in detailed. It is shown that both schemes reduce the sensitivity to the CSI errors, and the second scheme has a lower complexity.In summary, it is shown that the transceiver design schemes under different de-sign criteria and channel state information can significantly improve the performances of multi-user MIMO relay systems. Moreover, the proposed schemes can be applied to the wireless cellular networks, wireless local area networks, ad-hoc networks and wire-less sensor networks. The research results also provide the effective theoretical guidance for the combination of MIMO relay and other transmission techniques, such as cognitive radio, multi-cell, power transfer and so on.