Investigation Of Precoding In MIMO-OFDM Systems

In the fourth generation mobile communication scheme, OFDM and MIMO are regardedas one of the core technologies. Using orthogonal multi-subcarrier of OFDM, the systems canovercomes frequency selective fading wireless channel. Because the antennas at differentspatial positions can communicate simultaneously in MIMO, the wireless systems obtain aspatial diversity gain, enhance channel link quality and increase the channel capacity withmultiple bonuses. However, in order to combine OFDM and MIMO technology efficiently, itstill need to solve many problems of signal processing properly, including signalpreprocessing at transmitting node, signal detecting at receiving node, signal organization ofthe systems, the allocation of OFDM subcarriers, signal processing in relaying system and so.These issues are addressed in precoding research. Hence the precoding technology inMIMO-OFDM-based system is also very important to the new generation of wirelesscommunications technology.In this thesis, signal precoding algorithms are designed in MIMO-OFDM based system.The algorithms are classified according to the spatial diversity, spatial multiplexing,cooperative relaying precoding, multi-user technology, bidirectional transmission respectively.Different algorithm is applied to different scenarios with different systems. Some usefulattempts have been developed in the present thesis，including the following issues:First, for spatial diversity MIMO-OFDM downlink system, a double precodingalgorithm was designed including quantified beamforming vector and the quantified powervector of the spatial diversity scheme. First of all, according to the minimum total systemBER criterion, the beamforming vectors and power allocation coefficients were globallyjointly optimized. And the closed optimized solution was obtained as well. Then by usingbeamforming codebook and uniform distribution quantitative power codebook, the precodingvectors were quantified respectively. Finally, by using the correlation of adjacent subcarriers,the thesis proposed a clustering quantized feedback precoding and interpolation-recoveryalgorithm to further reduce the number of feedback bits. Simulation results show that, underfrequency-selective channel, the proposed algorithm has better overall performance than theexisting algorithm in terms of the amount of feedback bits, bit error rate characteristics. Second, for the spatial multiplexing MIMO-OFDM downlink system, a limited feedbackprecoding scheme was optimized as spatial multiplexing. First, the cost function was designedwith minimum mean square error. The transmitting precoding matrix and receiving decodingmatrix were optimized. And the power distribution was improve as well. Then, design apractical power quantization codebook and improve the power allocation in spatialmultiplexing systems. A subcarrier power allocation scheme was proposed with Lagrangemultiplier method. Finally, by using a suitable interpolation method spatial multiplexing,systems achieve a quantitative feedback under the relevant channel and precoding recovery.Simulation results show that in the case of OFDM subcarrier, the amount of feedbackalgorithms and error characteristics have better overall performance than the existingalgorithm.Third, for the three-node OFDM relay system which includes a source node, relay nodeand the destination node, we propose a multi-antenna limited feedback linear precodingscheme OFDM downlink. Algorithm was design by using the eigenvalue decomposition andQR decomposition of the channel matrix. The linear precoding algorithm systems compriseda S-D direct channel and S-R-D the channel. The source precoding matrix was independent ofserial number of OFDM subcarriers each frame, which can reduce the amount of feedbackbits. Finally,using the Lagrange multiplier method, according to the standards of maximizesum-rate in systems, propose a optimized subcarrier power allocation. Simulation resultsshow that the sum-rate of subcarrier channel is higher. The bit error rate is lower. And thefeedback bits is fitting for practical application.Fourth, for OFDM downlink multi-user relay system which comprised a single sourcenode,a single relay node and multiple destination nodes with S-D direct path, a linearprecoding scheme and convex optimizing precoding scheme are proposed. For linearprocessing scheme, based on the criterion of maximizingsignal-to-leakage-and-nose-ratio(SLNR), the precoding matrix were designed. The relayfunction was decomposed into S-R receiving and R-D transmitting stages. The algorithms arelower complexity. For convex optimization scheme, the optimum decoding matrix is designedthrough Wiener filter. Source node precoding matrix was obtained through using the objectivefunction of Weighted Sum-Rate and convex optimization method design. Solve the relay node precoding through the use of Daulity principles. Finally, in accordance with the criterionof maximizing the capacity of OFDM system, optimize the power allocation and subcarrierallocation of OFDM scheme. Simulation results and analysis show that the algorithm hasbetter characteristics in the error rate, the total capacity and the computational complexitythan the existing algorithm.Fifth, for multipoint to multipoint downlink MIMO-OFDM relay system, puts forwardtwo algorithms of multi-user precoding scheme. For the first algorithm, according to thecriterion of maximum rate in the system, construct the objective function containing only asingle variable under the introduction of weight matrix under minimizing the weighted MSEcriterion. The optimizing problem was transformed into a typical positive semidefiniterelaxation (SDR) And global optimization of precoding was obtained. For the secondalgorithm, we design a simple and practical scheme which only need to precode at relay node.Through the SVD decomposition of the channel, find the vector basis of channel zero spaceand design a relay node precoding matrix to achieve eliminating common channelinterference automatically. Finally, by introducing the objective function of maximizingchannel capacity under the source node and relay node power restriction, the subcarrierpower allocation were optimized using the KKT conditions of the convex function.Sixth, for bidirectional MIMO-OFDM relay system, design a full-duplex bidirectionaltransmission mode of precoding algorithm. First, proposes a precoding based on MSE summinimization criterion: The sum of signal MSE was decomposed into MA and BC tages. Inthe MA stage a gradient descent method was used to design the source precoding matrix. ForBC stage the relay precoding matrix was obtained by constructing dual item. And approachthe optimal matrix through iteration method. On this basis, design a user equalizationalgorithm based on the criterion of minimizing maximum MSE, by using theory of Schurcomplement,the optimizing problems was converted into a positive semidefinite(PSD) andresort to existing convex optimization soft package to solve the precoding optimizationproblems. Finally, using method of Lagrange dual function, solve the problem of theoptimized resources allocation of OFDM subcarrier. Analysis and simulation result show thatthe algorithm has better comprehensive performance in term of complexity, errorcharacteristics and channel capacity than the existing algorithm.