Browsing by Author "Nielsen, John"
Now showing 1 - 20 of 58
Results Per Page
Sort Options
- ItemOpen AccessA Millimeter-Wave Radar and Kinect Depth Camera for Non-Contact Respiratory Sensing(2013-04-04) Loblaw, Andrew; Okoniewski, Michal; Nielsen, JohnRemote, non-contact respiratory sensing has gained interest in the past decade with the dawn of cheaper microwave integrated circuits and optical devices. The need for an inexpensive and portable respiratory monitor is in particular demand in a hospital setting as the respiratory rate provides early warning for cardiorespiratory arrest. This thesis explores the hypothesis that an inexpensive millimeter-wave (MMW) radar module can synergize with an infrared (IR) camera to provide a low-cost module for accurate, long-term measurement of respiratory rate. Both the IR camera system and the MMW radar technique are experimentally validated. Promising results are presented for the IR camera in which the respiratory rate is accurately measured for different scenarios. Measurements with the MMW radar indicate that more expensive and complex hardware are required to achieve adequate results.
- ItemOpen AccessA modified fingerprinting technique for an indoor, range-free, localization system with dynamic radio map annealing over time(2012-09-13) Lesser, Andrew M.; Okoniewski, Michal; Nielsen, JohnIndoor wireless localization systems have gained considerable interest in the past decade with the wide spread implementation of affordable wireless networks throughout indoor environments. Many organizations have employed these systems to track people, equipment, and merchandise in an effort to reduce operating costs which can include loss or theft, inventory, and efficient utilization of time sensitive assets. The complex, indoor radio frequency propagation environment introduces many challenges for wireless location systems. In particular, the large and small scale fading of signals introduces uncertainties in the location dependence of radio frequency measurements. This thesis explores two approaches to mapping the above location dependency of measurements with the primary focus on reducing the time required for extensive environment calibration. The formulation of proposed location estimation algorithms and calibration approaches will be presented. A radio frequency device affixed to a mock hospital asset will be used as a real world example to validate the algorithms.
- ItemOpen AccessA step recovery diode based transmitter for UWB(2005) Shaskin, John Ernest; Nielsen, John; Davies, Robert J.
- ItemOpen AccessAn exploratory case study of educationally focused faculty development in engineering(2009) Clearwaters, Thomas; Nielsen, John; Kopp, Gail
- ItemOpen AccessAnalysis of Stroke Induced Motor Function Weakness in Post-Stroke Patients using Machine Learning(2021-09) Bhatt, Aakash; Yanushkevich, Svetlana; Souza, Roberto; Nielsen, JohnThe focus of this thesis is the development of a system that can analyze stroke induced motor weakness using pressure sensor mattresses. The proposed system utilizes time-series pressure data from publicly available datasets as well as data collected from recovering stroke patients at the Foothills Medical Center. Two tasks are performed with the pressure data. In the first task the incoming pressure data is classified into three body positions: supine, lateral, and prone on a frame-by-frame basis. The second task consists of classifying time-series pressure data into two classes: left-sided weakness and right-sided weakness. Results from the first task are used to improve results from the second task by only using patient data in which the patient is in a supine position. Extensive experiments are conducted using deep learning methodologies including convolutional neural networks and long-short term memory networks. The developed system is intended to be utilized to monitor patient condition throughout their stay at the hospital.
- ItemOpen AccessAutomated Pain Recognition Using Analysis of Facial Expressions(2017) Shier, Warren Adam; Yanushkevich, Svetlana; Nielsen, John; Shahbazi, MozhdehCurrent pain evaluation involves the use of patient self-reporting, which can be subjective, prone to suggestion, and infeasible on certain patients. Post-surgery patients, elderly people with dementia, or young children cannot properly convey their pain, even though it still occurs. There are also limitations on the frequency caregivers or doctors can check on their patients. To help solve this problem, this thesis develops solutions for automated pain detection via facial expressions. The system continually classifies the subject as being in pain, or not in pain. Subject pain levels are verified using the Prkachin and Solomon Pain Intensity Scale. Two fully automated algorithms are presented, the first uses Gabor filters with Support Vector Machines, the other uses a type of deep learning, Convolutional Neural Networks. Feasibility studies are conducted on a database and real-life subjects from an elderly care facility. Results are evaluated using precisions and speed of computation.
- ItemOpen AccessBiometric-Enabled Decision Support Platform with Risk Assessment(2022-01-14) Lai, Kenneth; Yanushkevich, Svetlana; Hatzinakos, Dimitrios; Hemmati, Hadi; Fear, Elise; Nielsen, JohnBiometric-based human trait and behavior recognition is a critical component of the rapidly growing domain of ambient intelligence. Particular applications of interest are biometric-enabled border checkpoints, access control, as well as healthcare and biomedical data analysis. In this thesis, we offer both theoretical and practical contributions. The main theoretical contributions include the framework for uncertainty measures and performance assessment measures in a decision support system. These measures include risk, trust, and bias and the methodology involves using these measures in a contemporary engine for decision support systems, based on causal models of uncertainty. These models allow for the prediction of events of interest and assess the risks associated with these events. Our main practical contributions include the advanced practical implementation of various machine learning approaches, mostly deep neural networks, to biometric-enabled applications such as facial recognition, action recognition, emotion classification, wearable data analysis for healthcare, and human-machine interaction applications. Demonstration of practical applications of machine reasoning for biometric-enabled systems that use facial recognition, action recognition, and emotion recognition is shown. In this research, we propose to combine multi-spectral biometric data processing, powerful deep learning techniques, along with performance improvement techniques, in a unified approach to automate face and action recognition. When combined, a platform consisting of powerful machine-learning techniques is used as a supporting tool to provide decision support for an operator. Multi-spectral data is understood as color and depth data such as video, depth, and derived skeleton joints. The deep learning techniques include convolutional and recurrent neural networks. The former is applied in our study to extract important spatial information from color and depth images, whereas the latter is utilized to recognize temporal patterns. Emerging deep learning model architectures are explored, one such network called the Residual Temporal Convolutional network offers improved performance in comparison to recurrent neural networks. This research will show the capability of using different types of data to train neural networks independently to recognize biometric patterns including actions and faces. Therefore, the main focus of this thesis is the development of a decision support platform for solving a variety of practical problems. These solutions are approached using the same methodology: advanced machine learning techniques are used to process, analyze and classify data, and machine reasoning is used as a common platform for the system-level decision making, with risk, bias, and/or trust assessment of the provided decision. This approach is embodied, in particular, in a proposed decision support system for human stress detection using physiological signals. Deep learning techniques were applied for detecting and recognizing different emotional states. The causal models were built upon the distribution of the detection and recognition scores in order to perform machine reasoning and information fusion. These results provided the operator with the risk assessment of any given scenario. Other examples of such systems are provided in multiple publications as reported in the thesis.
- ItemOpen AccessBlind calibration of ad hoc localization system(2011) Liu, Junjie; Nielsen, John
- ItemOpen AccessCompactly Supported Biorthogonal Wavelet Constructions on the A-star Lattice and their Application to Visualization(2017) Horacsek, Joshua; Alim, Usman; Alim, Usman; Samavati, Faramarz; Nielsen, John; Li, ZongpengIn this dissertation, a family of compact biorthogonal wavelet filter banks that are tailored to the geometry of the A-star lattice are derived. Our application of interest is on the three dimensional A-star or {\em body centered cubic} (BCC) lattice. While the BCC lattice has been shown to have superior approximation properties for volumetric data when compared to the Cartesian cubic (CC) lattice, there has been little work in the way of designing wavelet filter banks that respect the geometry of the BCC lattice. Since wavelets have applications in signal de-noising, compression, and sparse signal reconstruction, these filter banks are an important tool that addresses some of the scalability concerns presented by the BCC lattice. We use these filters in the context of volumetric data compression and reconstruction and qualitatively evaluate our results by rendering images of isosurfaces from compressed data.
- ItemOpen AccessComputer Vision based Indoor Navigation Utilizing Information from Planar Surfaces(2014-09-29) Dawar, Neha; Nielsen, John; Lachapelle, GérardTraditional wireless signalling based outdoor navigation techniques generally result in unsatisfactory performance for indoor environments due to low signal strength and multipath distortions. Computer vision (CV) sensors, due to their low cost and high performance, have gained enormous interest in indoor navigation over the past years. CV based 6DOF trajectory estimation is understood to be a computationally intensive ill-posed problem. Drastic simplification and enhanced robustness are possible in scenarios where camera observed features are constrained to a plane, such as a floor surface. Furthermore, if the features have geometric patterns, such as a regularly tiled surface, significantly more powerful constraints can be implemented. Exploration of such constraints is the aim of this thesis. Experimental results show that centimeter level accuracy in trajectory estimation can be achieved for arbitrary camera motion spanning several meters. As shown in this thesis, this accuracy is a result of constraints due to planar features observed.
- ItemOpen AccessContext aware high dynamics gnss-ins for interference mitigation(2011) Kamel, Ahmed Mohsen Mohamed; Lachapelle, Gérard; Nielsen, JohnAutonomous navigation systems used in missiles are mostly dependent on Global Positioning System (GPS) as a primary means of navigation. GPS usage has limitations in terms of missile high dynamics and expected signal interference in the battlefield. Due to the conflicting bandwidth requirements in Phase/ Frequency Lock Loops (PLLs/ FLLs), a novel FLL-assisted-PLL is proposed for very high dynamic conditions with reduced measurement noise and capabilities to cope with interference. The design is based on fuzzy systems and is used to generate the required Numerical Control Oscillator (NCO) tuning frequency with the information provided by phase and frequency discriminators. Detailed system design and performance analysis are presented where scenarios include high dynamics and different types and levels of interference are introduced. The designed system is compared also against 3rd order PLLs, with narrow and wide bandwidths, standard FLL-assisted-PLL, in addition to a Kalman Filter (KF) based PLL. Moreover, to insure robustness of the new system during periods of GPS blockage due to very high interference levels, the system is integrated with additional aiding that can provide external Doppler measurements. Two approaches are introduced to accomplish this. The first approach is to use the measurements available from the missile Inertial Measurement Unit (IMU) as these are not affected by interference. The second approach is to get the benefit from the modem GPS signals such as L2C signal which can be assumed not to be affected by interference if a narrow band interference signal is aimed at the primary GPS signal L 1. Performance assessments results demonstrate the enhanced performance of the proposed system where better tracking continuity during high dynamics up to 20 g's of acceleration and more accurate measurements in interference free environments and in low and medium interference levels up to jamming-to-signal ratios of 40 dB are achieved for the stand alone system, and at very high interference levels for the aided system.
- ItemOpen AccessDemocratizing Software Development and Machine Learning Using Low Code Applications(2022-12-19) Alamin, Md Abdullah Al; Uddin, Gias; Ruhe, Guenther; Far, Behrouz; Nielsen, JohnLow-code software development (LCSD) is an emerging approach to democratize traditional and Machine Learning (ML) application development for practitioners from diverse backgrounds. Traditional LCSD platforms promote rapid application development with a drag-and-drop interface and minimal programming by hand. Similarly, low-code Machine Learning (ML) solutions (aka, AutoML) aim to democratize ML development to domain experts by automating many repetitive tasks in the ML pipeline (e.g., data pre-processing, feature engineering, model design, and hyper-parameter configuration). The rapid emergence of LCSD platforms warrants systematic studies to understand the challenges developers/practitioners face while using the platforms. This thesis catalogs, for the first time in the literature, the challenges developers face while using low code platforms developed for traditional and ML software application development. To the end, we also offer our hands on experience of developing a low code ML software systems for our industrial partner. Specifically, we investigate the current status, i.e., services of LCSD providers, open-source research \& collaboration. We conduct the LCSD practitioners' challenges by analyzing their discussion on the popular Q&A forum Stack Overflow (SO) to seek technical assistance. To further validate our findings, we conduct to develop a low-code machine learning solution in collaboration with domain experts from industry and academia. Additionally, we develop AutoGeoML, an open-source low-code framework that solves the current limitations of low-code ML solutions. Our qualitative investigation of 121 traditional and 37 AutoML LCSD services shows that around 60% traditional LCSD solutions are related to business process management (BPM) and work process automation, 90% are proprietary, and only 63% platforms support only proprietary cloud deployment options. We find that around 65% of services offer shallow or general purpose ML applications, 57% solutions are open-sourced and offer flexible deployment options. According to our findings, Customization and LCSD Platform Adoption are the most discussed topic, followed by Data Management. We also find that Deployment and Maintenance are still the most difficult Software Development Life Cycle (SDLC) phase. We highlight the limitations of current low-code AutoML services and develop AutoGeoML. This open-source low-code framework provides domain-experts-in-the-loop customizability, and modular abstraction is some of the critical requirements lacking in existing AutoML frameworks. The findings of this thesis have implications for all three LCSD stakeholders: LCSD platform vendors, LCSD practitioners, and Researchers. Researchers and LCSD solution vendors can collaborate to improve different aspects of LCSD, such as better tutorial-based documentation, DevOps support, and expert-in-the-loop customizability.
- ItemOpen AccessDesign and implementation of a 5-channel CDMA receiver for mobile position location(2006) Lopez, Alfredo; Nielsen, John
- ItemOpen AccessDesign of Short Synchronization Codes for Use in Future GNSS System(2008-05-06) Shanmugam, Surendran K.; Mongrédien, Cécile; Nielsen, John; Lachapelle, GérardThe prolific growth in civilian GNSS market initiated the modernization of GPS and the GLONASS systems in addition to the potential deployment of Galileo and Compass GNSS system.The modernization efforts include numerous signal structure innovations to ensure better performances over legacy GNSS system. The adoption of secondary short synchronization codes is one among these innovations that play an important role in spectral separation, bit synchronization, and narrowband interference protection. In this paper, we present a short synchronization code design based on the optimization of judiciously selected performance criteria. The new synchronization codes were obtained for lengths up to 30 bits through exhaustive search and are characterized by optimal periodic correlation. More importantly, the presence of better synchronization codes over standardized GPS and Galileo codes corroborates the benefits and the need for short synchronization code design.
- ItemOpen AccessDesign of Short Synchronization Codes for Use in Future GNSS System(Hindawi Publishing Corporation, 2008) Shanmugam, Surendran K.; Mongrédien, Cécile; Nielsen, John; Lachapelle, Gérard
- ItemOpen AccessDevelopment of a general real-time multi channel IS-95 CDMA receiver for mobile position location(2008) Salimi, Nazila; Lachapelle, Gérard; Nielsen, John
- ItemOpen AccessDiscrete Fourier Transform Techniques to Improve Diagnosis Accuracy in Biomedical Applications(2018-01-08) Adibpour, Paniz; Smith, Michael; Fear, Elise; Frayne, Richard; Nielsen, JohnTransforming acquired data in time or space is necessary for many applications, due to practical constraints on time-domain sampling at high data rates or the requirement for algorithms to process frequency-domain data during the image reconstruction procedure. Therefore, the discrete Fourier transform (DFT) plays an important role in many fields for preprocessing, reconstruction or data analysis stages of algorithms. The hardware or physical constraints also necessitate acquisition of limited length raw data which results in DFT-imposed distortions after data processing for which low pass filters are considered as general solution. Through this thesis, fundamental DFT properties are investigated and an optimization method is introduced to take advantage of these properties. This method is a potential alternative to low pass filters which impose resolution loss to processed data. The formalized method is examined and validated using preliminary observer metrics for two magnetic resonance imaging reconstruction approaches and a microwave imaging technique.
- ItemOpen AccessDiversity Gain through Antenna Blocking(Hindawi Publishing Corporation, 2011-10-19) Dehghanian, Vahid; Nielsen, John; Lachapelle, Gérard
- ItemOpen AccessDynamic Eye-in-Hand Visual Servoing with Neural-Adaptive Backstepping(2019-03-22) Roy, Preston Logan George; Macnab, Chris J. B.; Nielsen, John; Goldsmith, Peter B.; Westwick, David T.This thesis investigates eye-in-hand visual servoing, where a camera on the robot arm provides information for the motor-control feedback loop. Current methods use a dual-loop strategy, where the outer-loop uses the visual servo error to compute desired joint velocities, while an inner-loop accomplishes the tracking. Since it is difficult to establish global stability with this strategy, this thesis instead investigates backstepping control. This provides a guarantee of uniformly ultimately bounded signals and explicitly accounts for the coupling between outer and inner loops. First a method with knowledge of the feature Jacobian is developed, then it is extended to an adaptive method that uses supervisory estimates of the feature Jacobian to maintain stable adaptation. The methods are further extended to the visual servo control of n-link robots, multiple features using a switched controller, and visual tracking. Non-linearities in the system are approximated using the computationally-efficient Cerebellar Model Articulation Controller neural network.
- ItemOpen AccessEnhanced cellular network positioning using space-time diversity(2007) Abdolhosseini Moghaddam, Ahmad Reza; Lachapelle, Gérard; Nielsen, John
- «
- 1 (current)
- 2
- 3
- »