Abstract
This doctoral thesis contains a collection of five papers preceded by an introduction. The papers investigate channel models for, design of, and performance analysis of wireless multiple-input multiple-output (MIMO) systems which are subject to a strong line-of-sight (LOS) channel component.
MIMO technology is embraced as one of the key technologies for fulfilling the demand for increased throughput and improved quality of service (QoS) in future wireless applications. This technology can both be employed to increase reliability, through diversity schemes such as e.g. maximum ratio combining and Alamouti coding, or to increase the spectral efficiency by spatial multiplexing schemes such as e.g. eigenmode transmission and V-BLAST. The performance of a wireless MIMO system is heavily dependent on the condition of the channel matrix, in the sense that the channel matrix should be of high rank for the MIMO system to achieve good performance. When the channel is such that the major part of the received power at the receiver (Rx) is due to multipath, fulfilling the high rank criteria is dependent on low correlation between the different subchannels. On the other hand, if the dominant component at the Rx is the deterministic LOS component, fulfilling the high rank criteria becomes dependent on the design of the two antenna arrays employed.
In this thesis we derive optimal antenna array designs for pure LOS channels with respect to mutual information (MI), when any combination of uniform linear arrays (ULAs) and uniform planar arrays (UPAs) are employed at the transmitter (Tx) and Rx. The important parameters with respect to design will be shown to be the antenna separation, antenna orientation, wavelength, transmission distance, and MIMO dimension. Moreover, we characterize the effects of these parameters deviating from their optimal values. The pure LOS channel matrix utilized is subsequently employed in a Ricean channel model also incorporating multipath, and performance is evaluated both analytically and numerically for different designs and multipath conditions. Furthermore, we investigate the performance of a possible future high frequency fixed wireless access (FWA) system based on the optimal design principle, to see how it works in a more realistic scenario. In general, the results show that a considerable gain is achieved if a design close to the optimal is possible for a MIMO system transmitting over a strong LOS channel.
The thesis also contains an analysis of the difference between the spherical wave model (SWM) and the plane wave model (PWM). The investigation is performed for systems utilizing ULAs, and it results in a framework that can be employed when evaluating when to apply the true SWM, and when the more simple approximate PWM gives sufficient modeling accuracy. Based on the framework, we conclude that the SWM should for example be applied in some typical WLAN scenarios.
Journal and conference contributions during
PhD studies. During the PhD studies, the author has contributed to the following journal and conference publications: