In the indoor environment, multipath fading causes the received signal amplitude to fluctuate rapidly over space and frequency. Multiple-in multiple-out (MIMO) systems overcome this phenomenon through the use of multiple antennas on transmitters and receivers. This establishes multiple independent MIMO sub-channels between antenna pairs, which allows a theoretical increase in capacity which is linear with the number of antennas, while requiring no additional power or bandwidth expenditure.The capacity increase is reliant upon MIMO sub-channels being well decorrelated. Decorrelation may be achieved by separating antennas in space. On devices where space is limited, an alternative approach is to use antennas with orthogonal polarisations, which may be positioned closer together. Existing literature states that the performance of polarisation MIMO systems is typically inferior to that of spatial MIMO systems under diversity applications, but can be superior in multiplexing applications. These statements are based on the analysis of a statistical channel model, using channel conditions assumed to be typical of an ideal polarisation MIMO system. There is little existing literature which examines how close these assumptions are to a practical polarisation MIMO channel, or whether the above statements remain true of practical systems.This thesis presents a novel end-to-end, predominantly deterministic approach to the modelling of polarisation MIMO systems. A bespoke MIMO channel model is used to estimate capacity and error rate under diversity and spatial multiplexing applications in the indoor environment. The parameters of the channel model are obtained deterministically from a ray launching propagation model, using antenna patterns of orthogonally polarised small antenna systems positioned in the indoor environment. The individual differences in the channel gains and K-factors of each sub-channel are accounted for. Correlation is accounted for using a full correlation matrix, rather than the Kronecker model. Particular attention is paid to mutual coupling of closely spaced antennas. Using this analysis, it is shown that for practical antennas and systems conditions of the polarisation MIMO channel may differ from those assumed in literature. The effect of this in terms of channel capacity and system bit error rate is directly determined and presented.Performance of polarisation MIMO systems, using co-located and spatially separate orthogonally polarised antennas, is compared to that of spatial MIMO systems, which use co-polar antennas with limited spatial separation. Additionally, comparison is made between compact polarisation MIMO systems which use orthogonal linear polarised antennas and those using orthogonal circular polarised antennas. Further analysis examines the significant effect of objects in the antenna near-field regions. The effects of the presence of a metal case on antenna performance are presented, before its impact on the channel conditions and ultimately the resultant MIMO performance is shown.
|Date of Award
|1 Aug 2017
- The University of Manchester
|Anthony Brown (Supervisor) & Ka Chun So (Supervisor)
- Ray tracing