Engineering Science - Theses, Dissertations, and other Required Graduate Degree Essays

Receive updates for this collection

Adaptive optics optical coherence tomography for in vivo retinal imaging

Author: 
Date created: 
2014-04-25
Abstract: 

Optical coherence tomography (OCT) is a non-invasive micrometer-resolution volumetric imaging modality that has been employed in diverse applications. In this thesis, we first describe a GPU accelerated program to perform FDOCT data processing and real time 3D volumetric rendering. The real time visualization of volumetric images provided by the GPU acceleration was essential to the rest of the work described in this thesis. Small animal models of retinal diseases serve as a vital component in vision research, and non-invasive in vivo imaging is becoming an increasingly important tool in the field. We describe the first adaptive optics optical coherence tomography (AOOCT) imaging system for high resolution mouse retinal imaging. Images of mouse retina acquired with AOOCT showed significant improvement in the brightness and contrast of capillaries and nerve fiber bundles. However, the accuracy of wavefront sensing limited the performance of AOOCT. A novel wavefront sensorless adaptive optics (WSAO) OCT system was developed to overcome the issues associated with conventional wavefront sensing. Combination of WSAO with OCT allows coherence gated, depth resolved aberration correction. Images of both pigmented and albino mouse retinas acquired using WSAO OCT system demonstrated superior image quality. The real time high resolution WSAO OCT was leveraged for a time course study of laser exposure in the retina.

Document type: 
Thesis
File(s): 
Senior supervisor: 
Marinko V. Sarunic
Department: 
Applied Sciences: School of Engineering Science
Thesis type: 
(Thesis) Ph.D.

Self-Organized Intelligent Distributed Antenna System in LTE

Author: 
Date created: 
2014-04-07
Abstract: 

In order to reduce the operational expenditure, while optimizing network efficiency and service quality, self-organizing network is introduced in long term evolution. The SON includes several functions, e.g. self-establishment of new base stations, load balancing, inter-cell interference coordination. Load balancing and inter-cell interference coordination are two of the most important self-organizing functions. In this thesis, load-balancing solution is investigated in order to optimize quality of service. To enable load balancing among distributed antenna modules, we dynamically allocate the remote antenna modules to the BTS sectors. Self-optimizing intelligent distributed antenna system is formulated as an optimization problem. Three evolutionary algorithms are proposed for optimization: genetic algorithm, estimation distribution algorithm, and particle swarm optimization. Computational results of different traffic scenarios after performing the algorithms, demonstrate that the the algorithms attain excellent key performance indicators. The downlink performance of cellular networks is known to be strongly limited by inter-cell interference in multi-carrier based systems when full frequency reuse is utilized. In order to mitigate this interference, a number of techniques have recently been proposed, e.g., the soft frequency reuse scheme. In this thesis, DAS is utilized to implement SFR. The central concept of this architecture is to distribute the antennas in a hexagonal cell such that the central antenna transmits the signal using entire frequency band while the remaining antennas utilize only a subset of the frequency bands based on a frequency reuse factor. A throughput-balancing scheme for DAS-SFR that optimizes cellular performance according to the geographic traffic distribution is also investigated in order to provide a high QoS. To enable throughput balancing among antenna modules, we dynamically change the antenna module's carrier power to manage the inter-cell interference. A downlink power self-optimization algorithm is proposed for the DAS-SFR system. The transmit powers are optimized in order to maximize the spectral efficiency of a DAS-SFR and maximize the number of satisfied users under different users' distributions. The results show that proposed algorithm is able to guarantee a high QoS that concentrates on the number of satisfied users as well as the capacity of satisfied users as the two KPIs.

Document type: 
Thesis
File(s): 
Senior supervisor: 
Shawn Patrick Stapleton
Department: 
Applied Sciences: School of Engineering Science
Thesis type: 
(Thesis) Ph.D.

Implementation and Evaluation of a QoS-aware Downlink Scheduling Algorithm for LTE Networks

Date created: 
2014-04-30
Abstract: 

Long Term Evolution (LTE) is becoming the mainstream of the fourth generation standard for high-speed wireless communications for mobile devices. Its radio access for downlink involves allocation of Physical Resource Blocks (PRB). In order to achieve optimal download performance for different applications to satisfy different QoS requirements, the downlink scheduling algorithm in use plays an important role in determining which PRBs and how are they allocated to each flow of bits. Several researches have exploited different scheduling strategies for flows; however, both the frequency and time domain allocations for PRBs should be taken into account. In this project, we implement and evaluate a QoS-aware downlink packet scheduling algorithm for LTE networks known as the Packet Prediction Mechanism (PPM) using the LTE Simulator (LTE-Sim). The PPM consists of three phases. It first utilizes the PRBs effectively in the frequency domain. It then manages queues and predicts the behaviour of future incoming packets based on the current ones in the queue by the concept of virtual queuing. Finally, it incorporates a cut-in process to rearrange the transmission order and discard overdue packets based on the predicted information from the previous phase. The simulation results demonstrate the effectiveness of the PPM scheme in achieving better downlink transmission performance in terms of Throughput, Delay, Fairness Index, Packet Loss Ratio (PLR), and Spectral Efficiency than other downlink schedulers such as Priority First (PF), Modified Largest Weighted Delay First (MLWDF), and Exponential Proportional Fair (EXPPF).

Document type: 
Graduating extended essay / Research project
File(s): 
Senior supervisor: 
Jie Liang
Department: 
Applied Sciences: School of Engineering Science
Thesis type: 
(Project) M.Eng.

Application of Narrow Band Power Line Communication in Volt/Var Optimization

Author: 
Date created: 
2014-03-21
Abstract: 

Smart grid is an advanced and sophisticated electrical network that uses a combination of information and communication technology to gather required data from different nodes on the network through a reliable, robust and cost effective communication solution, and acts on them to improve the efficiency and reliability of production and distribution of electricity. Volt/Var optimization (VVO) is one of the most important smart grid applications that gives this capability to utility companies to have an efficient electricity distribution network by maintaining an acceptable voltage level along the distribution section under different loading situations. This project focused on characterizing and evaluating the performance of narrow band power line communication (NB-PLC) as a prevalent communication technology for smart grid applications such as VVO in North American power grid. This work was done by establishing and setting up a real test bed in the lab. In this work we tried to implement S-parameter measurements due to its simplicity for channel characterization at high frequency ranges, when we have cascades of different components over the channel. It should be mentioned that the main focus of our work was on analyzing the behavior of an energized 5KVA MV/LV transformer over our channel. Then, we have shown the relationship between S-parameters and ABCD parameters to derive the channel transfer function based on ABCD matrix. Additionally, we did some simple noise measurements over our channel when it was non-energized and energized.

Document type: 
Graduating extended essay / Research project
File(s): 
Senior supervisor: 
Dr. Daniel Lee
Department: 
Applied Sciences: School of Engineering Science
Thesis type: 
(Project) M.Eng.

Investigation of Fluorescence Coherence Tomography for Optofluidic Applications

Author: 
Date created: 
2014-02-28
Abstract: 

Conventional Flow Cytometers (FC) and Fluorescence-Activated-Cell-Sorters (FACS) are mechanically complex, bulky, and require specialized human operators, large sample volumes, and sample preparation procedures for proper diagnosis of diseases such as leukemia and malaria. For this reason, there has been an increasing demand for miniaturization, reduction of cost, and portability of such devices. Lab-on-a-chip devices, which integrate microfluidics with other technologies, have been emerging as a potential solution to miniaturization of FC/FACS technology. One serious limitation of lab-on-a-chip devices is their inability to extract shape or morphological information which is very useful for cell differentiation and characterization. To meet this challenge, optical imaging techniques and microfluidics are combined to form a subset field in ‘optofluidics’. This thesis will help explore this field which describes systems that combine optics and microfluidics. In this thesis, as proof of principle, the integration of a novel optical imaging technique called Fluorescence Coherence Tomography (FCT) with microfluidics is presented. The FCT was used to measure the position of flowing fluorescent particles in the cross section of the microfluidic channel (perpendicular to the direction of flow). This type of measurement was motivated by recent reports in the literature demonstrating that a cell’s position in a microchannel is highly sensitive to its size and stiffness, which in turn are important biomarkers for cell classification. By combining FCT with microfluidics, the long term goal is to provide researchers and scientists with new possibilities for biological investigations in optofluidic applications. The preliminary results acquired through this work are important for future development of applications in the miniaturization of molecule specific flow cytometry.

Document type: 
Thesis
File(s): 
Senior supervisor: 
Marinko V. Sarunic
Department: 
Applied Sciences: School of Engineering Science
Thesis type: 
(Thesis) M.A.Sc.

Robotic single cell electroporation

Date created: 
2012-12-17
Abstract: 

Transporting DNA and metabolites across the cell membrane is a fundamental mechanism to achieve a controlled, quantitative understanding of the complex processes occurring in the human body at the cellular level. However, the thin, cell membrane forms an effective barrier to foreign molecules and alternative means are required to coerce entry. One method capable of overcoming this barrier is single cell electroporation (SCE) via microcapillary, which can be applied to cells directly in culture, or tissue. SCE induces reversible pores in the membrane by applying an electric field at the cell surface. Membrane-impermeable molecules enter these pores by electrophoresis and diffusion. The tip of the microcapillary can be fabricated with micrometer size geometries allowing extraordinary cell selectivity and access to small cell features with sparing quantities of molecules. However, the technical complexity of SCE limits the use to highly trained operators. Operators must carefully position a microcapillary tip on cells only several micrometers in height and must perform the technique using conventional microscopy methods that lack depth-perception. Additionally, knowledge of the electrical characteristics of SCE influencing the rate and efficiency are required. These broad technical requirements and the fragile nature of thin cell structures limit the efficiency of manual throughput. Furthermore, the sequence of tasks have not been adequately achieved by automated efforts, thus the true potential of SCE has not been realized. In this thesis, a versatile system and methods are described for an infrastructure designed completely for automated SCE. The intent of the system is to abstract the technical challenges and exploit the accuracy and repeatability of automated instrumentation leaving only the focus of the experimental design to the operator. In addition, new milestones within automated cell manipulation have been achieved. The system described herein has the capability of fully automated SCE of 'thin' cell features less than 10 m in height. This achievement eliminates limitations imposed by many mammalian cell lines and provides a rapid, transmembrane transport method for a broad range of applications. The execution is demonstrated by inserting a combination of a fluorescing dye and a plasmid DNA with a reporter gene into NIH/3T3 fibroblasts.

Document type: 
Thesis
File(s): 
Senior supervisor: 
Edward J Park
Department: 
Applied Sciences: School of Engineering Science
Thesis type: 
(Thesis) Ph.D.

Analysis and Design of Diversity Techniques for Terrestrial and Underwater Acoustic Communications

Author: 
Peer reviewed: 
No, item is not peer reviewed.
Date created: 
2013-12-20
Abstract: 

Since the early 1880’s, wireless broadband communications have been growing at explosive rates. While the personal communication systems have almost exhausted the spectrum, higher and higher data rates are required to support the ever demanding wireless services. Recently, to improve the spectral efficiency, diversity gains, and interference and power management for wireless multimedia and internet services, by combining the signals at both ends and effectively creating multiple parallel spatial data pipes, the multiple-input multiple-output (MIMO) technology has become a convenient framework. Motivated by these practical concerns, this thesis addresses the analysis and design of diversity techniques for terrestrial and underwater acoustic communication channels, in two parts. Part I studies novel relay selection strategies and diversity techniques for single carrier frequency domain equalization (SC-FDE) multi-relay cooperative networks, considering maximum-likelihood (ML) and minimum mean-square error (MMSE) receivers. We further extend our analysis to two-way relaying (TWR) networks, while incorporating different power control techniques. Building on our results on the diversity and error performance of the single relay and TWR cooperative systems, we extend our analysis to design of MMSE-based optimum beamforming matrices at user and relay terminals in a multi-user, multi-antenna TWR cooperative system. We further present a joint user-relay antenna selection algorithm by applying the estimation of distribution algorithm (EDA). The final contribution of the first part of this thesis is to extend our analysis to large relay networks and address the prohibitive computational and implementation complexity cost of the exhaustive search algorithms for joint transceiver/relay beamforming matrix design in large amplify-and-forward (AF) MIMO TWR networks, while incorporating the orthogonal matching pursuit (OMP) algorithm. The second part of this thesis focuses on the performance of differentially encoded space-time and space-frequency block coding techniques for terrestrial and underwater communication channels.

Document type: 
Thesis
File(s): 
Senior supervisor: 
Jie Liang
Sami Muhaidat
Department: 
Applied Sciences: School of Engineering Science
Thesis type: 
(Thesis) Ph.D.

Vital Signs Monitoring Using a New Flexible Polymer Integrated PPG Sensor

Author: 
Date created: 
2013-12-12
Abstract: 

For remote registration of vital signs, non-invasiveness is attractive. The possibility of acute loss of consciousness threatens the lives of employees such as soldiers, fire fighters, police officers, and law enforcement personnel, who perform duties at hazardous or remote conditions. The brain cells die within three minutes after occurrence of hypoxia in cerebrospinal fluid. Consequently, acute occurrence of cardiac or pulmonary system failure reduces an employee’s likelihood of surviving due to the lack of ability for communication and the lack of ability for emergency calls. In the case of severe injury, a medic’s response time becomes a crucial parameter for increasing a person’s likelihood of survival. Continuous monitoring of vital signs therefore assists medics and employees by reducing response time to severe accidents. For detection of vital signs; electrocardiography, capnometry and pulse oximetry are being widely used as the golden standard for extraction of vital signs such as heart rate, respiration rate and blood oxygenation. But, these sensors require motionless attachment to specific areas of the body. The fact that employees are constantly in motion and sometimes covered by protection shields introduces difficulty concerning the continuous obtaining of vital signs. In this dissertation, we studied feasibility of replacing Electrocardiography by Photoplethysmography on mechanically flexible sensors. Three major studies were performed. First, we developed an algorithm for the detection of respiration rate from Photoplethysmography. During this study, a fast respond CO2 sensor was also modulated to detect respiration rate, and both methods were compared to respiration effort transducer. Second, we assessed the feasibility of fabricating PPG sensors on plastic polymers. During this study, we designed and integrated a novel PPG device, using inkjet-printing technology. At last, we developed a computationally inexpensive algorithm for the extraction of heart rate variability (HRV) from the morphology of PPG. Results were compared to HRV from commercial ECG and the performance of the device was evaluated respectively.

Document type: 
Thesis
File(s): 
Senior supervisor: 
Bozena Kaminska
Department: 
Applied Sciences: School of Engineering Science
Thesis type: 
(Thesis) M.A.Sc.

Design and experimental proof of selected functions in implantable artificial kidney

Author: 
Date created: 
2013-12-12
Abstract: 

Renal failure results in poisoning because metabolic by-products are not promptly removed from the body. The main remedy for this condition is hemo-dialysis, where blood bypasses the kidneys and is filtered in a “dialyzer”, stationary machine. This research proposes and verifies novel techniques that allow an implantable device to replace a dialysis machine. This device would perform two important kidney functions: filtering solids and retaining desired electrolytes and small proteins. Three independent approaches are proposed and experimentally verified. The first approach is design optimization of the glomerular membrane as an implantable filter to separate blood cells from whole blood. We studied the parameters that minimized pressure drop per unit area in micro-channels (straight and diverging) with circular cross-sections. The second approach, aimed at extending the filtration capability of a porous membrane, used the concept of “back-wash”. It used a natural energy source in the body, the pulsatile character of blood flow, with pressure varying between 80 and 120 mm Hg. Under similar experimental conditions, experimental results demonstrated that the permeate volumetric flow rate was higher in the backwash system compared to the no-backwash system, and this flow rate could be maintained for many more filtration cycles. The third approach, which retained body electrolytes and small proteins, used a static electric field to divert blood ions and charged proteins back to circulation. Two geometries for this electrophoretic filtration were proposed and tested: “Y” method and “cross-flow” method. The “cross-flow” method seems more promising after a preliminary comparison. A benefit of using the electrostatic deviation of charged solids before mechanical filtration is a lower density of blood solids reaching the filtration membrane, causing a lower probability of filter clogging. Due to the importance of maintaining proper pressure drops at all renal filtration stages an implantable valveless pump was designed and fabricated for pressure drop adjustments. This pump’s novelty is that it relies entirely on blood pressure pulsations and does not require an external power supply. None of the proposed filtration techniques requires external power supplies; all rely on energy delivered by the heart.

Document type: 
Thesis
File(s): 
Senior supervisor: 
Dr Andrew Rawicz
Dr John Jones and Dr Will Cupples
Department: 
Applied Sciences: School of Engineering Science
Thesis type: 
(Dissertation) Ph.D.

Cognitive Radio: Energy Sensing Analysis and Dynamic Spectrum Access modeling

Author: 
Date created: 
2013-11-25
Abstract: 

Efficient utilization of the spectrum has become a fundamental requirement in modern wireless networks, due mainly to spectrum scarcity and the ever-increasing demand for higher data rate applications and internet services. A particularly interesting proposal to meet this requirement is the cognitive radio (CR) system which can adapt its transmission parameters according to the environment. CRs, as will be shown in later chapters, are very efficient in maximizing spectrum utilization due to their inherent spectrum sensing capability.The purpose of this dissertation is to investigate and analyze two main components of CR. First is the sensing or exploring component, which is the core of a CR device as it is the first stage to discover spectrum holes (SHs) in a spectrum band. For this component, a new algorithm to compute the detection probability in the case of odd degrees of freedom and a closed-form expression for the detection probability in Nakagami-m fading channels are presented, both for a local spectrum sensing scenario. For a cooperative scenario, the errors of CRs decisions which are caused by erroneous feedback channels are analyzed. In addition, the optimal number of CRs that are required to mitigate against such errors is derived. The second component is the access or exploiting component, i.e. how a CR device can exploit SHs efficiently. To study the second component, the interactions between the primary users (PUs) and secondary users (SUs) are modeled as a continuous time Markov chain (CTMC). Based on the CTMC model, the effect of two inevitable sensing errors (misdetection and false alarm) on the blocked call probability, the dropped call probability and system utilization is investigated for two access schemes. In the first scheme, the PUs are considered to access the system using a standard access policy. In the second scheme, the PUs use non-standard access policies. In both schemes, the overall (primary and secondary) system utilization is analyzed and compared under both perfect and imperfect sensing. The simulation results obtained concur with the analytical ones and it is determined that spectrum utilization can be improved by choosing a suitable non-standard access policy.

Document type: 
Thesis
File(s): 
Senior supervisor: 
Jie Liang
Department: 
Applied Sciences: School of Engineering Science
Thesis type: 
(Thesis) Ph.D.