Research On The Multiple Antenna Channel Modeling And Simulation Techniques For The Next Generation Of Wireless Communication Systems

As the popularization of smart phones and the exponential growth of wire-less data demands, the next generation of wireless communication systems need faster information transmission rate, higher energy utilization efficiency and better heterogeneous networks integration. Designing and optimizing these sys-tems, therefore, necessitates better understanding on the characteristics, mod-els and emulation approaches of wireless propagation channels. In this the-sis, the major topic is the modeling and emulation techniques of multiple-input multiple-output (MIMO) channels in the next generation of wireless communi-cation systems. The innovations include:1) A new algorithm for estimating the characteristic parameters of spatial multi-paths is proposed. The widely used algorithm for multipath pa-rameter estimation is the space alternating generalized expectation-maximization (SAGE). The SAGE algorithm has the disadvantages of enormous calculations and possible convergence to local optimums. The precision of estimated param-eters may decrease significantly especially when the directional array is used in measurements. In this thesis, some analyses on angular resolution of MIMO arrays are performed, and based on these analyses, the particle swarm optimiza-tion (PSO) algorithm is utilized to improve optimum searching ability. The overall algorithm that proposed is called iterative least square particle swarm optimization (ILS-PSO). It not only simplifies complexity and guarantees the global optimum, but also supports antenna grouping in measurements and the combined optimum search between groups, which may significantly improve the multipath tracking ability in high mobility environments. In both numeri- cal simulations and measurements post-processing, the results from ILS-PSO is validated to match realistic environment better than the results from SAGE, while the amount of calculations is only1/4of that in SAGE. It also supports multipath parameter estimation in high mobility channels.2) From the perspective of channel capacity, the information carry-ing ability of massive MIMO channels is analyzed. Firstly, the subspace compactness is defined to quantitatively describe the relationship between the MIMO channel capacity and the power angular spectrum of multipath channel, and the mathematical connection from MIMO channel singular values and sub-space compactness is derived. Based on these results, the favorable propagation condition (FPC), which maximizes the capacity of massive MIMO systems, is considered. The existence of FPC in the scenarios where UTs are co-located is studied under a simplified geometric model. A precise closed-form expres-sion and an approximated simplified expression for asymptotic FPC are derived. From numerical simulations, it is validated that these mathematical expressions are good approximations for the existence of FPC in the realistic environment when the number of array elements is quite large. This conclusion can be used in budgeting needed number of array elements in massive MIMO systems.3) A differential channel quantization algorithm is proposed, and fur-thermore, power allocation issues in multi-user MIMO (MU-MIMO) sys-tems are studied. Non-coherent trellis coded quantization (NTCQ) is a chan-nel quantization algorithm for massive MIMO systems. It has low complexity, and is easy to be implemented. However, the strong space-time correlation in massive MIMO channels is not well utilized in NTCQ. In this thesis, based on NTCQ, a differential quantization is proposed, in which rotating and skewing operations in the codebook space are performed to adaptively track the varying space-time correlation, and only the update of channel state information (CSI) is quantized. This may further improve the quantization efficiency while keep-ing the complexity low. Numerical results reveal that compared to NTCQ, the proposed algorithm can improve transmission quality in both massive MIMO systems and ordinary MU-MIMO systems. Based on the quantized CSI at trans-mitter (CSIT), co-channel interference is further studied, and a self-convergent power allocation algorithm is proposed. The upper bounds for the SINRs at UTs are further derived. Numerical results illustrates the block error rate (BLER) improvement and capacity gain that obtained from proposed power allocation.4) Issues on radio-frequency (RF) level channel emulation, including the system architecture and key techniques, are discussed. The channel em-ulator (CE) is an equipment that connects the realistic RF-ends of both the BS and the UT and emulates the realistic propagation channels. In this thesis, the system architecture and software flow paths of a CE are firstly designed. Then, some solutions to the key technologies in CEs are proposed. As to the fractional delay digital filtering technique, which is used to precisely simulate the path de-lays, based on the Farrow structure, a scheme that consumes constant logical resources while providing a flexible configuration on the delay parameter is designed. The amplitude error and the phase error of this scheme are below10-3. As to the MIMO over-the-air (MIMO-OTA) testing, which is used to reproduce spatial multipath environment in anechoic chambers, firstly an ideal scheme that is able to provide precise reproduction is designed. Then, a low cost scheme that is able to reproduce spatial correlations is described. It can re-produce spatial correlation for any2-elements UT using an8-elements chamber array with error below10-2.