Investigating the Effect of Antenna Polarization on the Performance of CoMP Systems based on Synchronous Multi-link Measurements

Abstract: In this master thesis work, the effect of polarization, at the Base Station (BS) side, on the performance of multi-user Coordinated Multipoint (CoMP) systems is studied. This study was performed using synchronous multi-link measurements that took place at the campus of the faculty of Engineering, LTH, Lund University, Lund, Sweden, where two different BS setups were studied. In the first setup, one BS provided with one antenna array consisting of four antenna elements, was used. The antenna aperture size was varied from 0.17 m to 24 m, where different antenna polarizations (single- and cross-polarized arrangements) were considered. In the second setup, we use two BSs located 60 m apart, each of which is provided with two co-located antenna elements spaced by half a wavelength. Two antenna polarizations at the BS are studied: single- and cross-polarized arrangements. In both setups, four virtual users, spaced 0.5 m apart with one cross-polarized antenna were considered. For each setup, the user Multiple Input Multiple Output (MIMO) channels are used to evaluate the sum-rate capacity of the system, where the minimum mean square error (MMSE) beamforming at both the BS and the Mobile Station (MS) was used. Furthermore, in the second setup, the influence of user hand and torso into the MS antenna patterns and hence into the resulting performance was incorporated. For the first setup, i.e., using one BS antenna with variable aperture, it was found that: 1) using cross-polarized antenna elements at the BS improves the sum-rate capacity by about 35% and 72% in Non Line of Sight (NLOS) and Line of Sight (LOS), respectively, if the aperture of the antenna array is small (less than 1 m). 2) Increasing the BS array aperture gives better sum-rate capacity to a certain point, then the improvement saturates. 3) If the BS array aperture is "large enough", then the performance improvement gained from using BS cross-polarized antennas is insignificant compared to using single-polarized ones. For the second setup, i.e., using two BSs each of which is provided with co-located antennas with half a wavelength inter-element spacing, it was found that: the cross-polarized antenna configuration improves the ergodic sum-rate capacity about 50% compared to the single polarization configuration. In addition, increasing the number of antenna (from 1 to 2) at the MS side yielded an improvement of 43% in sum-rate capacity.