Research On MIMO Technology For The Next Generation Wireless Communication

Multiple-Input-Multiple-Output (MIMO), as a key technique in wirelesscommunication, can fully explore the potential of space resources by transceivinginformation using more than one antenna, and it has the advantage of high spectralefficiency and large capacity. And massive MIMO technique plays a significant role in thefifth-generation mobile communications system (5G), for it can dramastically improve thespectral efficiency applying very large number of antennas. Besides, massive MIMO canfocus the transmitting and receiving energy in a very small region to improve systemenergy efficiency. Therefore, this work aims at studying MIMO technology for the nextgeneration wireless communication and the main contributions are as follows:Firstly, a robust OFDM-MIMO transceiver design method is proposed which aims atminimizing mean square error (MSE) of signal detection under the total transmitting powerconstraint. The solution of our proposed algorithm is similar to the traditional water-fillingmethod. However, it has varying water-filling bottoms since we have taken the channelestimation error into account and we call it “cluster water-filling”. The cluster water-fillingsolution still has a closed form bearing much practical use.Then a robust MIMO precoder design method under channel uncertainty is providedwith a capacity maximizing criteria. Considering channel estimation error, we formulate aconvex MIMO capacity maximizing problem with transmitting power constraint andcompute out the approximate optimal solution using cluster water-filling algorithm. And theproposed method is simulated and proved to have better performance compared with theexisted non-robust method.Afer that, a semi-blind channel estimation method for multi-cell multi-user massiveMIMO systems is studied. The proposed method takes advantage of the channel vectororthogonality between different users in massive MIMO systems and estimates the channelusing data signals and a few pilot signals based on eigen-value decomposition (EVD)method. This proposed method solves the much pilot consumption for channel estimationproblem in massive MIMO systems and performs well in the simulation analysis part. Finally, we investigate the joint optimization method for pilot and data signal resourceallocation. The proposed method gives the optimal signal structure based on MMSE criteriaand demonstrates the relationship between achievable sum rates and the pilot signal lengthas well as pilot power allocation. Then a convex optimization method is applied to obtainthe optimal solution of the formulated problem. At last, the performance is assessed bysimulation results.