Effects of mutual coupling between the antennas on the cellular mimo systems

For current and next generation wireless communication networks, there are critically essential system requirements such as having more robust performances and higher data rates. By introducing multiple antennas at both transmitter and receiver, Multiple-Input Multiple-Output (MIMO) systems, in rich scattering wireless environments, channel capacity may increase as well investigated in the literature. The role of the spatial diversity in MIMO systems is quite significant and it is ideally based on the statistically uncorrelated channels from each other in the space. Spatial diversity is one of the significant techniques for MIMO communication systems. It is especially being applied to combat the difficulties of detection of the receiving signal because of multipath fading environment. By multiplying the antennas, it is desired to combine different multiple signals at the receiver and it is used efficiently to fight against the fading in modern wireless communication systems. Spatial correlation itself is the key parameter of the spatial diversity which determines the system performance in MIMO wireless communication systems in terms of the quality of service and theoretical channel capacity as investigated before. There are many studies that explore the channel spatial correlation in various propagation environments and power angular spectrum which is the statistical distribution of received power regarding the angle of arrival. One of this distributions and the most well-known one is uniform distributed power angular spectrum in two dimensional plane for Rayleigh fading channels. Solution of the spatial correlation under the uniform distributed power angular spectrum between the two arbitrary points in the two dimensional space is equal to zero-th order Bessel function of the first kind as a function of the distance between the points, and this popular spatial correlation model is known as Jakes' Model. For rich-scattering Rayleigh fading channels, this correlation model is being used often-times in the literature. When the antennas are located closely to each other, voltage and current distributions are going to be varied with existing of the other antenna(s) and this will eventually make change on the input impedances and far-field radiation characteristics of the antennas. This mutual coupling effect between the antennas should also be considered on calculating the spatial correlation as well. It is again well investigated that the effects of mutual coupling between the antennas for both transmit and receive side for MIMO systems. However, studies which are specific to antennas in cellular communication systems are very rare. Studies that investigate the effect of mutual coupling on spatial correlation are mostly based on the dipole antennas located at both transmitter and receiver sides can be considered as impractical for cellular communication systems. It is strongly needed to do similar studies with practical cellular communication devices and realistic antennas to expanding the literature. Aim of this study is investigating the effects of mutual coupling on the spatial correlation and the system performance with the antennas designed specifically for a cellular communication system that modelled in the environment has two dimensional uniform distributed power angular spectrum which is known as Jakes' Model. Besides, it is also aimed to compare the calculated spatial correlations in this study with the measured spatial correlation in an example multipath fading environment, as well. This thesis consists of two main studies in order to investigate the mutual coupling effects between the antennas on spatial correlation and channel capacities. First study is a calculation the spatial correlation through simulation that performed on electromagnetic simulation program CST Microwave Studio. This study is can be considered as a pre-study regarding to second study which consists of both simulation and realization of antenna designs on commercial smart-phone. Both studies have their own results in terms of the effects of mutual coupling between the antennas designed. First part of the study is checking the mutual coupling effects with more realistic antennas for both receive and transmit side. Unlike the studies in the literature, different antenna types are considered. For transmitter side (considered as a base station), half-wavelength, "lambda / 200" radius thin wire dipole antennas and for receiver side (considered as a mobile station), a wire inverted-F antennas with a ground plane on FR-4 substrate whose dimensions are similar to practical smart-phone case were designed. FDD-LTE Band 7 (2600 MHz) is considered to model the proposed system and both dipoles and inverted-F antennas are designed to radiate at 2600 MHz band. Mutual coupling effect on spatial correlation is investigated through 2x2 cellular MIMO communication system includes two dipole antennas at transmitter and a one hand-held device that has two wire inverted-F antennas (IFA) at receiver side. To investigate the mutual coupling effect on the spatial correlation of the proposed system, spatial correlations and ergodic channel capacities were computed for several conditions in terms of the spacing between the antennas at both receiver and transmitter independently. To consider the mutual coupling effect on spatial correlation, a coupling matrix which is defined in detail is calculated through electromagnetic simulation and introduced as a modification matrix to correlation matrix. Then, the original spatial correlation matrix which is formed on Jakes' Model is modified with this coupling matrix and created a new correlation matrix which includes mutual coupling effect. To compute the ergodic capacity, 50,000 i.i.d channel realizations have been generated and calculated for several situations as fixing antenna spacing values as practical as d = 1.2 lambda for mobile station and d = 3 lambda for base station. For all 50,000 channel realizations, channel matrices are modified with spatial correlation matrices that includes mutual coupling effects. It is observed through the numerical simulations that the mutual coupling reduces the spatial correlation for both antenna types, however it may increase or decrease the channel capacity for some antenna spacings. Second part of the study is on the investigating the mutual coupling effects by modelling the receiver side as a commercial smart-phone. To investigate the mutual coupling effect, 4 inverted-F antennas were designed and realized on the inner plastic rear cover of the smart-phone with adhesive copper tape. All the antennas designed for FDD-LTE Band 7, like the pre-study. For simulation of the antennas, detailed three-dimensional model of the smart-phone was used. Simulation and measurement results which includes far-field gain measurements in anechoic chamber were compared. Designed 4 antennas are grouped as a MIMO pairs which are considered to use in the receiver side of a 2x2 MIMO system. For all 6 antenna pair combinations, distances between the antennas are 14.5 mm, 125 mm, 130 mm, 127.8 mm, 125.8 mm and 58.6 mm; which are equal to 0.128 lambda, 1.106 lambda, 1.15 lambda, 1.131 lambda, 1.113 lambda and 0.518 lambda respectively in terms of free-space wavelength at 2655 MHz that is corresponds to mid-channel of the downlink frequency band of FDD-LTE Band 7. Mutual coupling matrices of all possible antenna combinations for receiver side of 2x2 MIMO system, spatial correlations for Jakes' Model and mutual coupling effect included correlations were calculated and compared with simulation results as well by using the same method with the first study. Ergodic channel capacities were also calculated for the all possible 6 combinations of designed 4 inverted-F antennas. Unlike the first study, some antenna couples have larger spatial correlation after the mutual coupling effect is included to the spatial correlation calculation. Another finding of this study is channel capacities for all antenna combinations have been increased by adding the mutual coupling effect to Jakes' Model. Besides, to calculate the receive spatial correlations of the designed antennas in an example multipath fading channel, VNA-based channel measurements were performed in an office has a dimensions of 3.5 m x 3.5 m x 3.5 m by changing the position of a designed quarter wavelength monopole 70 times which is considered as a transmitter antenna. To observing the difference on spatial correlation, two antenna couples has been selected that predicted to be have high and low spatial correlations. Spatial correlations of two receiver antenna combinations were computed by using 70 different measurements and the obtained result was compared with the previous study whose spatial correlation calculation includes the mutual coupling effect with respect to Jakes' Model. Results showed that, the spatial correlations were consistent with previous study regarding to have proportional spatial correlation values. To conclude, in order to investigate mutual coupling effects between the antennas in a cellular MIMO system, which is the main goal of this thesis, two different studies has been done. First study is an examination of the mutual coupling effect to spatial correlation and channel capacity in terms of the antenna spacings for both transmitter and receiver with different antenna types. What is determined in this study is while spatial correlations are always reducing by the effect of mutual coupling, channel capacities are fluctuating with different spacings. Second study is a similar one with the first study, but with 4 antennas with fixed positions on a commercial smart-phone. Results for this study is partly different in terms of spatial correlation which is calculated as having larger values than the correlation based on Jakes' Model for several antenna pairs. However, computed channel capacities have been observed as tending to increase by adding mutual coupling effect. In addition to second study, VNA-based channel correlation measurements in an example multipath fading environment have been performed to comparing with the calculated spatial correlation. Final measurement verified the second study as having rational values as expected.