Browsing by Author "Potter, Michael E."
Now showing 1 - 12 of 12
Results Per Page
Sort Options
- ItemOpen AccessBlind Compensation of Impairments in Wireless Transceivers(2018-07-24) Aziz, Mohsin; Ghannouchi, Fadhel M.; Helaoui, Mohamed; Fapojuwo, Abraham Olatunji; Potter, Michael E.; Sawan, MohamadModern wireless communication systems suffer from hardware imperfections that degrade the quality of transmission signals and make the detection of signal quite difficult at the receiver. This thesis focuses on the gain and phase imbalances caused by the modulators and demodulators and nonlinearity stemming from the transmitter power amplifier. Broadly, the contribution of this thesis is two folds: Blind solutions to mitigate the above mentioned hardware impairments of the wireless link through the proposal of a methodology based on the derivation of closed form expressions for the probability density functions (PDFs) of the signals in the presence of these impairments. In this regards, firstly, a PDF in the presence of modulator’s and demodulator’s in-phase and quadrature phase imbalances has been derived and validated. A maximum likelihood estimation of the imbalance parameters has been proposed to mitigate these imperfections. The proposed methodology has been evaluated using extensive simulations and measurements. To evaluate the static performance of the proposed methodology, 10 KHz modulated signal has been used. Measurement results show that an image rejection of greater than 30 dB can be achieved. For a larger bandwidth signal of 1 MHz, around 19 dB improvement in NMSE can be achieved using the proposed methodology, as compared to the uncompensated case. Secondly, a closed form PDF in the presence of gain and phase imbalances and the transmitter’s power amplifier nonlinearity has been derived and validated. A cumulative distribution function-based methodology has been adopted to mitigate the effects of power amplifier’s amplitude distortions. For the modulator’s impairments, a maximum likelihood estimation of the imbalance parameters has been used to estimate and compensate for the modulator’s imperfections. Using measurements, for a 3 MHz LTE signal, a normalized mean squared error and an error vector magnitude of -35 dB and 1.5% can be achieved, respectively.
- ItemOpen AccessDesign of mems capacitive switches for a tunable reflectarray element(2008) Wu, Billy; Okoniewski, Michal; Potter, Michael E.
- ItemOpen AccessDetection algorithms for an alternative mode of high-density optical data storage(2006) Goss, Kelly Christine; Potter, Michael E.
- ItemOpen AccessDetection systems and algorithms for multiplexed quantum dots(2012) Goss, Kelly Christine; Potter, Michael E.; Messier, Geoffrey G.Quantum Dots (QDs) are semiconductor nanocrystals that absorb light and re-emit at a wavelength dependent on its size and shape. A group of quantum dots can be designed to have a unique spectral emission by varying the size of the quantum dots (wavelength) and number of quantum dots (optical power) [l]. This technology is refered to as Multiplexed Quantum Dots (MxQD) and when it was first proposed, MxQD tags were created with 6 optical power levels and one QD colour or 3 QD colours and 2 optical power levels. It was hypothesized that a realistic limit to the number of tags would be a system of 6 optical power levels and 6 QD colours resulting in 46655 unique tags. In recent work, the fabrication and detection of 9 unique tags [2] was demonstrated which is still far from the predicted capability of the technology. The limitations affecting the large number of unique tags are both the fabrication methods and the data detection algorithms used to read the spectral emissions. This thesis makes contributions toward improving the data detection algorithms for MxQD tags. To accomplish this, a communications system model is developed that includes the inteference between QD colours, Inter-Symbol Interference (ISI), and additive noise. The model is developed for the two optical detectors, namely a Charge-Coupled Device (CCD) spectrometer and photodiode detectors. The model also includes an analytical expression for the Signal-to-Noise Ratio (S R) of the detectors. For the CCD spectrometer, this model is verified with an experimental prototype. With the models in place, communications systems tools are applied that overcome both ISI and noise. This is an improvement over previous work in the field that only considered algorithms to overcome the ISI or noise separately. Specifically, this thesis outlines the proposal of a matched filter to improve SNR, a Minimum Mean Square Error (MMSE) equalizer that mitigates ISI in the presence of noise and a Maximum Likelihood Sequence (MLS) detection algorithm to provide the best detection performance at the cost of higher computational complexity. Simulations demonstrate that the improved data detection algorithms could read 46,655 MxQD tags with over 97% accuracy compared to 11 % or 60% with traditional data detection algorithms for the CCD spectrometer and photo diode detector respectively
- ItemOpen AccessFrequency tunable double negative materials(2006) Darling, Michael; Okoniewski, Michal; Potter, Michael E.
- ItemOpen AccessImplementation of General Dispersive Anisotropic Materials in Lebedev FDTD(IEEE, 2018-09-17) Salmasi, Mahbod; Potter, Michael E.; Okoniewski, Michal M.A variant of the finite-difference time-domain (FDTD) algorithm - the Lebedev FDTD method - is augmented to allow for the simulation of electromagnetic waves in general dispersive anisotropic materials. Distinct dispersions may be applied to each of the principal axes of the anisotropic medium. The method is based on the auxiliary differential equation (ADE) and is simple to implement. Because of the collocated nature of the algorithm, multiterm dispersions necessary in previous implementations are avoided. Lebedev grid has previously been studied in detail by different research groups for simulation of anisotropy. In this work, we specifically focus on extending frequency dispersive materials to this method, and provide multiple examples to demonstrate the accuracy, validity and stability of the method.
- ItemOpen AccessImproved efficiency of marine em subsurface modeling using the total-field scattered-field fdtd formulation(2012) Dukeshire, Andrea; Potter, Michael E.; Okoniewski, MichalTo efficiently model the effect of the airwave in the controlled-source electromagnet-1cs (CSEM) method has historically been a problem of major interest. The methods proposed in this thesis provide a new method of simulating CSEM interactions in the finite-difference time-domain (FDTD) by using the total-field scattered-field (TFSF) formulation. TFSF provides an inherent separation of the total and scattered fields, which allows the proposed method to decompose the CSEM domain into separate simulation runs. By simulating the problem in stages, the process of extracting the response of the scatterer is streamlined. If the increased efficiency options within the proposed method are utilized, the efficiency of single forward simulations can be improved. This method shows potential for computational cost savings, especially in typical CSEM data analysis and reconstruction situations where numerous iterations of various scatterers are required.
- ItemOpen AccessInvestigation on a hertzian dipole antenna above an artificial impedance surface(2008) Zhou, Mi; Potter, Michael E.; Okoniewski, Michal
- ItemOpen AccessOn the nature of discrete plan wave and huygens' source for a perfectly matched tfsf implementation in the fdtd scheme(2010) Tan, Teng Meng; Potter, Michael E.Scattered wave problems resulting from objects exhibiting complex geometries and high heterogeneities which often occur in applied science or engineering applications can only be solved by numerical means. An important class of these problems comes from scatterers being radiated by a far-field source such as a plane wave. In the FDTD (Finite-Difference Time-Domain) numerical method, the Total-Field Scattered-Field (TFSF) technique is a Huygens' source algorithm that can source this far-field problem efficiently. For most applications the time-domain IFA (Incident Field Array) method has been successfully used for a number of years. Other applications such as radar-cross-section (RCS) evaluation, backscattered field from biomedical cells, or buried objects where the material contrast against the uniform background is small, can all have a scattered field well below the incident wave, i.e. a dynamic range smaller than -100 dB. An accurate frequency domain known as the AFP ( Analytic Field Propagator) method was recently developed to address such a small scattering analysis. However, the preprocessing stage necessitates prodigious data storage, therefore limiting t his accurate technique to 2D and a special case of 3D problems. This thesis first introduces a one-to-many mapping technique to reduce a general 3D plane wave source into a 1D representation. Hence, the memory storage problem of the AFP method is mitigated significantly. It is also demonstrated that an FDTD plane wave propagation direction can be expressed in terms of a discrete angle called the rational angle. The one-to-many mapping and rational angle are then used to construct a true time-domain numerical plane wave source valid for any scheme where partial derivative operators are approximated with finite differencing counterparts; the technique developed here is not unique to FDTD algorithms. When used as the Huygens' source for a 1D / 2D / 3D TFSF formulation, the proposed algorithm completely removes the spurious scattered field associated with existing formulations. This solution has never before available to the FDTD community. Because of the recursive algorithm, the large memory issue is also avoided. Numerical examples confirm the leakage error is at the finite precision limit (-300 dB for double precision).
- ItemOpen AccessPerformance Evaluation and Improved Permittivity Estimation for a Transmission-based Microwave Imaging System(2019-09-19) Owjimehr, Mehri; Fear, Elise C.; Potter, Michael E.; Belostotski, LeonidMedical microwave imaging, especially breast imaging, is one of the interesting applications of microwave imaging. This imaging technique has potential as a new breast imaging modality as it is low-power, not expensive, and doesn't require tissue compression and inconvenience. Several microwave breast imaging techniques have been introduced and show both promise and limitations. The system of interest in this thesis, the transmission system, appears to have the ability to overcome some of those limitations. Although the transmission system doesn't generate high quality images, it shows reasonable estimation of the average permittivity of the scanned tissue. Phantom and human scans using this system show potential, leading to further investigation. In this thesis, the performance of the transmission system is evaluated in terms of accuracy and resolution. Then one of the biggest challenges in image reconstruction, which is multipath, is investigated. The information extracted from multipath is used to improve the average permittivity estimation and form cross-sectional maps. The proposed methods are also tested on volunteer data and show reasonable results in terms of consistency and tracking changes. Finally, an improved denoising algorithm is proposed to reduce the background noise of the reflected signals in order to prepare them for average permittivity estimation.
- ItemOpen AccessScheduling algorithms for cooperative communication networks(2009) Lambert, Nathalie; Messier, Geoffrey G.; Potter, Michael E.
- ItemOpen AccessSynergistic Combination of Microwaves and Acoustic Signals: Towards Improved Breast Imaging(2019-03-01) Omer, Muhammad; Fear, Elise C.; Potter, Michael E.; Sesay, Abu B.; Ferguson, Robert J.; Yu, Alfred C. H.Microwave-based breast imaging techniques offer the convenience of frequently-repeated scans involving no breast compression, non-ionizing radiation, inexpensive scan costs, sensitivity to water content and low-complexity 3D image reconstruction. These benefits are well-suited for its application in breast health monitoring and treatment surveillance for which there is no current established approach. However, limited imaging resolution and target localization accuracy necessitate improvements in image quality for clinical applications. The integration of prior structural information has been seen to benefit microwave-based breast imaging techniques, including microwave tomography (MWT) and microwave radar (MWR)-based approaches. The sources of this structural prior explored in the literature include alternate imaging modalities such as computed tomography (CT) and magnetic resonance imaging (MRI). The ultrasound (US) signals with their short wavelength, low loss and small footprint may also provide this opportunity. The aim of this work is to investigate the possibility of extracting prior information from US signals and integrating this information into radar-based microwave image reconstructions. The synergistic combination of these two modalities improves the imaging focus, resolution, target detection and localization accuracy. In addition, the integration of structural information provides important contextual information for results interpretation as well as enhances the robustness of the method to the variations in breast shape, size, density, number of tumour inclusions and their locations. To support this development process, methods of constructing and generating a repository of realistic numerical breast models have also been presented. These models offer flexibility and reconfigurability to meet varying requirements of developing imaging technologies. They are applied to develop the proposed dual-modality imaging technology, verify its performance and compare it with the state-of-the-art MWR imaging approach.