A Generalized Approach to Model and Optimize Multi-Input Multi-Output Communication and Imaging Systems
Open Access
- Author:
- Fadlullah, Jarir
- Graduate Program:
- Electrical Engineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- October 01, 2010
- Committee Members:
- Mohsen Kavehrad, Dissertation Advisor/Co-Advisor
Mohsen Kavehrad, Committee Chair/Co-Chair
Shizhuo Yin, Committee Member
James Kenneth Breakall, Committee Member
Md Amanul Haque, Committee Member - Keywords:
- MIMO
communication systems
imaging
turbulence
copper cables
equalization
interference cancellation
clouds
monte carlo ray tracing
infrared
visible light communications - Abstract:
- The use of multiple elements in radio frequency (RF) wireless communication systems has been proven to be more robust to rich scattering environments, which otherwise contributes to random fading, long considered detrimental to traditional singleelement systems. However, performance improvements obtained by multiple-input multiple-output (MIMO) systems can be severely limited by physical constraints such as antenna separation, which directly affects the random statistics with correlation. In this dissertation, we review various determining factors and their effects on a MIMO RF wireless system and apply the conclusions to other areas for modeling and optimization. Until now, MIMO systems in communications have been restricted to wireless mediums. However, the basic tenets of MIMO transmission also apply to a general communication or imaging system. This is evident in copper cable Ethernet transmission, where the user has access to four parallel twisted-pair wires and the performance is limited by inter-channel-interference (ICI), not only inter-symbol-interference (ISI). These interferences are usually mitigated by employing single-input single-output (SISO) equalizers and interference cancellers. However, in a MIMO model, the cross-channel interferences carry useful information that can be salvaged by employing a MIMO equalizer/canceller, and thus contribute to signal-to-noise-ratio (SNR) increment. We investigate the effectiveness of different MIMO equalization and cancellation techniques in such semi-static channels, and provide a comparison of SNR and bit-error-rate (BER). Signal constellation design for modulation and coding is also an important factor for achieving the theoretical capacity bounds, and these are presented for hypothetical systems supporting 10Gbps, 40Gbps and 100Gbps data rates. The mathematical analysis for MIMO systems easily extends itself to other communication systems, such as free space optical (FSO) links. These communication links suffer from log-normal fading due to scintillation prevalent in distances longer than a hundred meters. Cloud and fog also impose barriers of attenuation in these channels. A simulation model is used in conjunction with scintillation characteristics to show the benefits of using multiple transmitters and receivers for FSO links. Furthermore, MIMO techniques are extended to indoor optical wireless communications in the infra-red wavelength range, which have the potential to provide data rates in the gigabits per second range. A MIMO modeling technique is presented here, which gives an accurate representation of the channel, and MIMO optimization techniques can be employed which use the available channel resources judiciously to improve data rate and error performances. Active optical imaging is a parallel to communication systems, with spatial information added on to the temporal information, where the latter is the only concern in communications. The mathematics behind MIMO has the flexibility to incorporate multiple variables, in this case space and time, easily into analysis and simulation. Active optical imaging by means of laser is a well-known technology used by remote sensing and surveillance communities; however, these suffer tremendously in the presence of scattering particles and turbulence. Scintillation is the greatest obstacle in imaging cases, and it causes the ideal point spread function (PSF) to broaden and distort or blur the received image. MIMO techniques, comparable to MIMO equalization for communication systems, can be applied to blindly deconvolve the received image from the distorted PSF. These novel techniques are discussed in this thesis, along with contrast and resolution improvements they offer. The contributions of this thesis are manifold. The concept of MIMO has been applied to several communication and imaging systems and is liberated from the RF wireless communications paradigm, and meaningful quantitative and qualitative connections have been drawn to applications other than RF wireless, namely FSO and indoor optical communications, Ethernet copper cable transmission, and deblurring for active imaging. It is expected that the value of the findings will trigger a tremendous motivation in thinking ‘jointly’ in terms of multiple variables rather than in terms of the single time or space-dependent variables which communication engineers are inclined to favor. The value of the analyses also lies in the fact that communications and imaging are shown to be similar problems, when cast into a general MIMO framework.