Computer Engineering

Permanent URI for this collection

https://hdl.handle.net/10315/30135

Browse

Now showing 1 - 20 of 24

Open Access
Coordinated Sensor-Based Area Coverage and Cooperative Localization of a Heterogeneous Fleet of Autonomous Surface Vessels (ASVs)
(2015-08-28) Forooshani, Parisa Mojiri; Jenkin, Michael R.
Sensor coverage with fleets of robots is a complex task requiring solutions to localization, communication, navigation and basic sensor coverage. Sensor coverage of large areas is a problem that occurs in a variety of different environments from terrestrial to aerial to aquatic. In this thesis we consider the aquatic version of the problem. Given a known aquatic environment and collection of aquatic surface vehicles with known kinematic and dynamic constraints, how can a fleet of vehicles be deployed to provide sensor coverage of the surface of the body of water? Rather than considering this problem in general, in this work we consider the problem given a specific fleet consisting of one very well equipped robot aided by a number of smaller, less well equipped devices that must operate in close proximity to the main robot. A boustrophedon decomposition algorithm is developed that incorporates the motion, sensing and communication constraints imposed by the autonomous fleet. Solving the coverage problem leads to a localization/communication problem. A critical problem for a group of autonomous vehicles is ensuring that the collection operates within a common reference frame. Here we consider the problem of localizing a heterogenous collection of aquatic surface vessels within a global reference frame. We assume that one vessel -- the mother robot -- has access to global position data of high accuracy, while the other vessels -- the child robots -- utilize limited onboard sensors and sophisticated sensors on board the mother robot to localize themselves. This thesis provides details of the design of the elements of the heterogeneous fleet including the sensors and sensing algorithms along with the communication strategy used to localize all elements of the fleet within a global reference frame. Details of the robot platforms to be used in implementing a solution are also described. Simulation of the approach is used to demonstrate the effectiveness of the algorithm, and the algorithm and its components are evaluated using a fleet of ASVs.
Open Access
Attention and Sensor Planning in Autonomous Robotic Visual Search
(2015-08-28) Rasouli, Amir; Tsotsos, John K.
This thesis is concerned with the incorporation of saliency in visual search and the development of sensor planning strategies for visual search. The saliency model is a mixture of two schemes that extracts visual clues regarding the structure of the environment and object specific features. The sensor planning methods, namely Greedy Search with Constraint (GSC), Extended Greedy Search (EGS) and Dynamic Look Ahead Search (DLAS) are approximations to the optimal solution for the problem of object search, as extensions to the work of Yiming Ye. Experiments were conducted to evaluate the proposed methods. They show that by using saliency in search a performance improvement up to 75% is attainable in terms of number of actions taken to complete the search. As for the planning strategies, the GSC algorithm achieved the highest detection rate and the best efficiency in terms of cost it incurs to explore every percentage of an environment.
Open Access
Distributed State Estimation For Smarter Electric Power Grids
(2015-08-28) Saxena, Shivam Kumar; Asif, Amir
The focus of this thesis is to design and implement distributed and decentralized state estimation (SE) algorithms for smart Electric Power Grids (EPGs). These algorithms are applied to two different types of EPGs: 1) modern, deregulated transmission networks that include advanced wide-area monitoring systems, and; 2) smart distribution networks with high penetration of distributed and renewable generation (DG) configured of microgrids. Microgrids are capable of cutting off from the main grid and operating autonomously in the islanded mode of operation in case of emergency situations. SE in such systems is complex because of highly non-linear system dynamics, slow and corrupted measurement updates, as well as the sheer computational complexity of the estimation algorithms. The contribution of this thesis is to explore the design and implementation of a reduced-order, distributed particle filter for state estimation in EPGs. Knowledge of the EPG state is necessary for EPG control, optimization, and emergency troubleshooting.
Open Access
Compressive Sensing and Time Reversal Beamforming Approaches for Ultrasound Imaging
(2015-12-16) Hantira, Noura; Asif, Amir
The objective of this thesis is to develop a novel beamforming technique for ultrasound machines that enables field reconstruction at sampling rates much lower than the Nyquist rate. In our simulations, we use Field II, a MATLAB based program for simulating transducer fields and models of biological tissues for imaging applications. Field II is capable of generating the emitted and pulse-echo fields for a large number of transducers configurations, including linear, circular, and rectangular arrays. Once the ultrasound field is determined, the proposed imaging technique is applied to the received signals to reconstruct the image for reference biological tissues. Applying different adaptive beamforming techniques, including the delay and sum (DAS) and Capon algorithms, the received signals from Field II simulation program are used to render the ultrasound images. A second goal of the thesis is to apply compressive sensing (CS) on received signals to reconstruct full-length signals from a reduced number of samples. A third goal is to couple the principal of time reversal (TR) with compressive sensing to extend the CAPON beamformer for reconstructing images of biological tissues at low sampling frequencies in rich multipath environments. The outputs of compressive sensing and CAPON-based algorithms, alone or in conjunction with each other, are severely degraded in such environments. Through numerical simulations, I illustrate an enhancement in reconstructed quality of images depicting biological tissues with my time-reversal based compressive sensing, CAPON approach.
Open Access
Towards CMOS Nuclear Magnetic Resonance Spectroscopy: Design, Implementation and Experimental Results
(2016-09-20) Pourmodheji, Hossein; Ghafar-Zadeh, Ebrahim
Nuclear Magnetic Resonance (NMR) Spectroscopy is used intensively along with other ancillary spectroscopic and characterization techniques. The design and implementation of High Throughput NMR Spectroscopy is a key challenge to accelerate the drug discovery process. On the other hand, the current conventional NMR technologies are expensive and bulky. The development of novel handheld NMR spectroscopy is a key challenge towards NMR spectroscopy for Point-of-Care (PoC) diagnostics applications. This thesis addresses the above-mentioned challenges of High Throughput NMR Spectroscopy and Handheld NMR spectroscopy by developing new integrated circuits dedicated to NMR spectroscopy using Complementary Metal Oxide Semiconductor (CMOS) technology. Simulation and characterization results were also used to prove the functionality and applicability of the proposed techniques. We have designed two CMOS chips using 0.13-m technology, first chip includes number of new vertical microcoils and LNA with 780 pV/Hz at 300 MHz and the second one is a new dual-path NMR receiver.
Open Access
Automatic Speed Control For Navigation in 3D Virtual Environment
(2016-11-25) Papoi, Domokos Mihaly; Stuerzlinger, Wolfgang
As technology progresses, the scale and complexity of 3D virtual environments can also increase proportionally. This leads to multiscale virtual environments, which are environments that contain groups of objects with extremely unequal levels of scale. Ideally the user should be able to navigate such environments efficiently and robustly. Yet, most previous methods to automatically control the speed of navigation do not generalize well to environments with widely varying scales. I present an improved method to automatically control the navigation speed of the user in 3D virtual environments. The main benefit of my approach is that automatically adapts the navigation speed in multi-scale environments in a manner that enables efficient navigation with maximum freedom, while still avoiding collisions. The results of a usability tests show a significant reduction in the completion time for a multi-scale navigation task.
Open Access
Voltage and Reactive Power Control in Islanded Microgrids
(2017-07-27) El-Taweel, Nader Ashraf Abd El-Fattah; Farag, Hany E. Z.
Previous studies put on view lots of advantages and concerns for islanded microgrids (IMGs), whether it is initiated for emergency, intentionally planned or permanent island system purposes. From the concerns that have not been addressed yet, such as: 1) The ability of the distributed generation (DG) units to maintain equal reactive power sharing in a distribution system; 2) The ability of the DG units to maintain acceptable voltage boundary in the entire IMG; 3) The functionality of the existing voltage and reactive power (Volt/Var) DG, this thesis analyzes the complexity of voltage regulations in droop-controlled IMGs. A new multi-agent algorithm is proposed to satisfy the reactive power sharing and the voltage regulation requirements of IMGs. Also, the operation conflicts between DG units and Volt/Var controllers, such as shunt capacitors (SCs) and load-ratio control transformer (LRT) during the IMG mode of operation, are investigated in this thesis. Further, a new local control scheme for SCs and LRTs has been proposed to mitigate their operational challenges in IMGs.
Open Access
Shape Perception of Clear Water in Photo-Realistic Images
(2017-07-27) Sultana, Arhum; Allison, Robert
Light plays a vital role in the perception of transparency, depth and shape of liquids. The perception of the surfaces of liquids is made possible with an understanding of refraction of light and knowledge of the underlying texture geometry. Given this, what specific characteristics of the natural optical environment are essential to the perception of transparent liquids, specifically with respect to efficiency and realism? In this thesis, a light path triangulation method for the recovery of transparent surface shape and a system to estimate the perceived shape of any arbitrary-shaped object with a refractive surface are proposed. A psycho-physical experiment was conducted to investigate this using the perceived shape of water from stereo images using a real time stereoscopic 3-D depth gauge. The results suggest that people are able to consistently perceive shape of liquids from photo-realistic images and that regularity in underlying texture facilitates human judgement of surface shape.
Open Access
Embedded CMOS Basecalling for Nanopore DNA Sequencing
(2017-07-27) Wang, Chengjie; Magierowski, Sebastian; Ghafar-Zadeh, Ebrahim
DNA sequencing is undergoing a profound evolution into a mobile technology. Unfortunately the effort needed to process the data emerging from this new sequencing technology requires a compute power only available to traditional desktop or cloud-based machines. To empower the full potential of portable DNA solutions a means of efficiently carrying out their computing needs in an embedded format will certainly be required. This thesis presents the design of a custom fixed-point VLSI hardware implementation of an HMM-based multi-channel DNA sequence processor. A 4096 state (6-mer nanopore sensor) basecalling architecture is designed in a 32-nm CMOS technology with the ability to process 1 million DNA base pairs per second per channel. Over a 100 mm^2 silicon footprint the design could process the equivalent of one human genome every 30 seconds at a power consumption of around 5 W.
Open Access
Performance Analysis of Coexistence Schemes for LTE in Unlicensed Bands
(2017-07-27) Jabbariardebili, Lida; Eckford, Andrew W.
LTE in the unlicensed spectrum, is becoming a popular area of research. Since LTE-Unlicensed (LTE-U) provides subscribers with higher-quality mobile voice, and video experience in high-traffic or low-signal locations, a fair coexistence mechanism with other networks, like Wi-Fi is essential. In this thesis, we propose two coexistence mechanisms that could be employed to ensure a fair channel access. First, we consider coexistence mechanism fundamentals, and then downlink system performance of two coexistence mechanisms are analyzed for multi-operator LTE-Unlicensed (LTE-U) deployments with different simulation scenarios, using NS-3. First we introduce the most trustworthy coexistence mechanism, and then a high-performance coexistence scenario is provided. We conclude that Licensed Assisted Access (LAA) can coexist with Wi-Fi without impacting Wi-Fi more than an equivalent Wi-Fi network. In the second part, uplink performance evaluation of LTE in licensed spectrum is also demonstrated.
Open Access
An Empirical Assessment on the Techniques Used in Load Testing
(2017-07-27) Gao, Ruoyu; Jiang, Zhen Ming
There are two main problems associated with load testing research: (1) the testing environment might not be realistic and (2) lack of empirical research. To address the first problem, we systematically assess the performance behavior of the system with various realistic environment changes. Results show that environment changes can have a clear performance impact on the system. Different scenarios react differently to the changes in the computing resources. When predicting the performance of the system under new environment changes, our ensemble-based models significantly out-perform the baseline models. To address the second problem, we have empirically evaluated 23 test analysis techniques. We have found all the evaluated techniques can effectively build performance models using data from both buggy or non-buggy tests and flag the performance deviations. It is more cost-effective to train models using two recent previous tests collected under longer sampling intervals.
Open Access
Nonnegative Code Division Multiple Access Techniques in Molecular Communication
(2017-07-27) Wang, Linchen; Eckford, Andrew W.
In molecular communication, two types of multiple access have been studied: time division and molecule division. In this work, we consider code division multiple access. However, unlike code division multiple access that has been used for electromagnetic signals, we investigate optical code division multiple access: since molecular signals have the same non-negativity feature as optical signals, this scheme is a promising solution for molecular communication. In this thesis, we perform experiments and set up simulation models which match these experiments. Moreover, using simulations, we find the features of optical code division multiple access for molecular communication. Our results include an optimal information transmission scheme, and an algorithm to decode molecular information signals. Finally, we demonstrate reliable communication with multiple access by using this scheme.
Open Access
Characterizing and Improving Logging Practices in Java-based Open Source Software Projects - A Large-scale Case Study in Apache Software Foundation
(2017-07-27) Chen, Boyuan; Jiang, Zhen Ming
Log messages (generated by logging code) contain rich information about the runtime behavior of software systems. Although more logging code can provide more context of the system's behavior, it is undesirable to include too much logging code. Yuan et al. performed the first empirical study on characterizing the logging. In the first part of the thesis, we conduct a large-scale replication study on characterizing the logging practices on Java-based open source projects. A significantly higher portion of log updates are for enhancing the quality rather than co-changes with feature implementations. However, there are no well-defined coding guidelines for performing effective logging. In the second part, we studied the problem of characterizing and detecting the anti-patterns in the logging code. We have encoded these anti-patterns into a static code analysis tool, LCAnalyzer. Case studies show that LCAnalyzer has an average recall of 95% and precision of 60% .
Open Access
RTOS Control of Hardware Processes
(2017-07-28) Atiwa, Sumaia; Aboelaze, Mokhtar
In this thesis, adding hardware-process support to Microcontroller Real-time Operating System Version 2 (MicroC/OS-II) is proposed. MicroC/OS-II is a hard real-time operating system (RTOS), mostly written in the C programming language. MicroC/OS-II is designed to manage limited resources within embedded systems, and it can only execute and control software processes performed in the same processor system. MicroC/OS-II has been modified in order to manage external hardware processes. These hardware processes are implemented on a Nexys 3 Spartan-6 FPGA Board. In this thesis, MicroC/OS-II is already ported to run on an EVBplus HCS12 development board with CodeWarrior Embedded Software Development Tools from Freescale Semiconductor Inc. Modifications are applied on MicroC/OS-II interrupt system to manage hardware processes, and SPI protocol and parallel interface are set up to communicate between the HCS12 trainer and the FPGA board. The work is illustrated by designing a satellite attitude controller, using variable structure control (VSC).
Open Access
What's Missing in your Shopping Cart? A Set Based Recommendation Method for "Cold-Start" Prediction
(2018-03-01) Zhou, Yubo; Xu, Jia
This thesis studies the problem of predicting the missing items in the current user's session when there is no additional side information available. Many recommender systems fail in general to provide a precise set of recommendations to users with limited interaction history. This issue is regarded as the "Cold Start" problem and is typically resolved by switching to content-based approaches which require additional information. In this thesis, we use a dimensionality reduction algorithm, Word2Vec under the framework of Collaborative Filtering to tackle the "Cold Start" problem using only implicit data . We have named this combined method: Embedded Collaborative Filtering ECF. We are able to show that the ECF approach outperforms other popular state-of-the-art approaches in "Cold Start" scenarios by 2-10% regarding recommendation precision. In the experiment, we also show that the proposed method is 10 times faster in generating recommendations comparing to the Collaborative Filtering baseline method.
Open Access
Autonomous Trail Following
(2018-05-28) Sefid, Masoud Hoveidar; Jenkin, Michael R.
Trails typically lack standard markers that characterize roadways. Nevertheless, trails are useful for off-road navigation. Here, trail following problem is approached by identifying the deviation of the robot from the heading direction of the trail by fine-tuning a pre-trained Inception-V3 [1] network. Key questions considered in this work include the required number, nature and geometry of the cameras and how trail types -- encoded in pre-existing maps -- can be exploited in addressing this task. Through evaluation of representative image datasets and on-robot testing we found: (i) that although a single camera cannot estimate angular deviation from the heading direction, but it can reliably detect that the robot is, or is not, following the trail; (ii) that two cameras pointing towards the left and the right can be used to estimate heading reliably within a differential framework; (iii) that trail nature is a useful tool for training networks for different trail types.
Open Access
Implementation Of A Novel Cooperative Protocol for Distributed Voltage Control in Active Distribution Networks
(2018-08-27) Qamar, Rabbia; Aboelaze, Mokhtar
Microgrids are small localized grids that help to integrate many renewable-energy sources into the main electric grid. Microgrids can also operate separately from the main electric grid during faults to enhance the customers reliability. For a successful integration of microgrids we need to control the voltage at the distributed generation units in order to achieve the required sharing of reactive power. For this purpose a multiagent based distributed control scheme is implemented in this thesis. The objective of this thesis is to design and implement a multiagent system for the microgrid that has distributed battery energy storage systems (BESS) and renewable distributed generation (DG) units. The proposed multiagent system has been designed to coordinate among distributed generation (DG) units to control voltage. Multiagent system is composed of multiple agents that communicate to solve problems. The proposed multiagent system for the control of microgrid has been implemented on Texas Instruments Tiva-C controller boards. The real time simulator Opal-RT has been used to create a microgrid model. Hardware testing is done in real time.
Open Access
A Quasi-Resonant Bidirectional Converter with Soft-Switching Operation for Energy Storage Applications
(2018-08-27) Aggarwal, Devina; Lam, John
The increased penetration of renewable energy power systems to produce clean and sustainable energy has led to the increased usage of various types of energy storage devices, such as high power density battery technologies, flywheel energy storage and super-capacitors. Energy storage devices are essential in any renewable generation systems to ensure providing uninterruptible and reliable power. Typically, a power electronic converter is required to serve as the intermediary between the common grid in a renewable energy system and the energy storage device. To be specific, the power converter must be able to facilitate bidirectional power flow between the grid and the energy storage device. Since the voltage level of the energy storage device is often much lower than the grid voltage level, the bidirectional converter must ensure that the voltage level can be stepped up or down efficiently as per the system requirements depending on the direction of the power flow. In this thesis, a unique quasi-resonant bidirectional converter topology is proposed for energy storage application. The proposed circuit only requires two switches to achieve bidirectional power flow. Hence, compared to the conventional dual-active bridge (DAB) based bidirectional converter topologies that require 8 switches, the total number of active switching devices required the proposed topology is greatly reduced. In addition, both switches in the proposed topology are able to achieve zero voltage switching (ZVS) turn-on and zero current switching (ZCS) turn-off to minimize the switching power losses without using additional auxiliary circuits. The operating principles and design equations of the proposed circuit will be discussed in details in this thesis. An extended version of the proposed topology that employs a modular design structure for high power application is also presented and discussed. Simulation results and experimental works on a proof-of-concept hardware prototype are given to highlight the performance of the proposed bidirectional converter.
Open Access
Operation of Over-Current Relays During Double-Phase Operation Mode of Power Distribution Systems
(2018-11-21) Hajiseyedolia, Seyedahmad; Hooshyar, Ali
Power system reliability is one of the objectives of every electric utility. Statistical analyses of faults in power systems demonstrate that most of the faults are asymmetrical, most commonly, single-phase to ground. Despite conventional breakers, modern breakers are capable of disconnecting the faulted phase(s). Disconnection of the faulted phase(s) increases the reliability of the system. However, a disconnected faulty phase can cause different challenges for distribution system. These challenges for distribution systems adversely affect the operation of protective devices. In this research, the effects of operating the distribution system with a disconnected phase double-phase operation mode (DPOM)on operation of over-current relays are analyzed. A new method for setting the relays properly during DPOM is presented. In this method, two type of settings, called "prompt settings" and "optimum settings" are presented. The methodology evaluation results prove that the method is capable of re-coordinating the relays during DPOM.
Open Access
Subjective Assessment of Image Compression Artefacts on Stereoscopic Display
(2019-03-05) Mohona, Sanjida Sharmin; Allison, Robert
Image and video quality are important to depict any pictorial information vividly and correctly. With the advancement of technology, we can produce high-quality images and can display those in advanced high-resolution displays. But as high-quality images continue to increase in size, transmitting these exceeds the limited bandwidth of display links. To cope, we need to compress the images but desire that the user cannot perceive any difference between the compressed and uncompressed images. In my thesis, psychophysical experiments with a flicker paradigm were undertaken to do a subjective assessment of the visibility of compression artefacts of two sets of images with two codecs viewed on a stereoscopic display. For one set of images the result shows that artefacts can be silenced in some stereo images relative to 2D while testing with the other set of images was inconclusive. This thesis documented evidence for silencing of artefacts in 3D displays. Other differences between stereoscopic and 2D presentation can be predicted but were not observed here (perhaps due to floor effects). Further large-scale subjective assessment with challenging images may help to get a concrete conclusion.