LURA 2019

Posted on

<<

The Fourth annual Lassonde Undergraduate Summer Student Research Conference (LURA) showcasing summer research students currently working on projects with faculty from the Lassonde School of Engineering will be taking place on Thursday, August 15, 2019. 

The conference is a full-day event filled with oral presentations and poster sessions where over 70 students will display the research they’ve been working hard on all summer. There is something for everyone with innovative projects ranging from 3D Bone Marrow Engineering, the Effects of Green Walls on Air Quality, and Human-Computer Interaction in Virtual Reality.

Come support the bright minds of Lassonde by RSVPing here.

Meet some of the researchers and their projects below:

Austin Martins-Robalino, Department of Civil Engineering
Research Program: NSERC USRA
Supervisor: Dan Palermo

Austin Martins-Robalino is a 4th-year student in the Department of Civil Engineering. He spent the summer exploring the effects of emerging specialty concrete composites reinforced with smooth reinforcement on structural behaviour in Dr. Dan Palermo’s laboratory.

By summer’s end, Austin hopes to gain a better understanding of the overall improvement of bond strength of smooth bars utilizing UHPFRC and ECC. This is important due to increased use of smooth reinforcement, in the form of shape memory alloys, in critical structures to increase resilience to seismic events.

Abstract:

An Investigation of Utilizing Smooth Reinforcement in UHPC and ECC and Impact on Flexural Behaviour

The aim of the research is to strengthen the understanding of how use of smooth reinforcement in reinforced high-performance concrete beams affects the overall flexural behaviour of the structural element. With the emergence of Shape Memory Alloys (SMAs), which are smooth reinforcing bars, as a promising material to increase the seismic resilience of structures, the differences in structural behaviour compared to conventional deformed steel reinforcement must be quantified. As part of these comparisons, the research involved designing 12 beams constructed from conventional normal-strength concrete, Engineered Cementitious Concrete (ECC), and Ultra High-Performance Concrete (UHPC), with the conventional concrete as the control and the latter two materials being used in structures sensitive to seismic activity. Each material consisted a set of 4 beams which varied in reinforcement type, smooth or deformed, with hooked or straight ends. In order to determine the flexural strength, beams were tested under four-point bending using a universal testing machine. Midpoint deflection was simultaneously measured using a linear potentiometer and digital image correlation. In addition, beams reinforced with straight bars involved the measurement of slip of each reinforcing bar. As a complementary study, the beams were also modelled in a finite element analysis program, VecTor2, to determine which theoretical models best predict the observed experimental results. These results can provide insight into the merit of using SMAs as reinforcement in high strength concretes to overcome the loss of bond-slip strength inherent with the use smooth reinforcement. Such insight would prove valuable to the future of improving seismic design and understanding of such a promising material.

 

Ivan Mishev, Department of Earth & Space Science & Engineering
Research Program: LURA
Supervisor: Sunil Bisnath

Ivan Mishev is a 3rd-year Earth and Atmospheric Science student at York University. Having specialized in Geomatic Science, Ivan is spending the summer exploring Global Navigation Satellite Systems (GNSS) in Dr. Sunil Bisnath's GNSS Laboratory. 

Specifically, Ivan will be working with a GNSS signal simulator to develop output files that can be used with Precise Point Positioning (PPP) GNSS measurement processing algorithms. This mechanization will enable the GNSS Lab to assess PPP algorithms more thoroughly and allow for more accurate error modelling. 
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
The aim of this research is to improve accuracy from 10s of metres to the centimetre-level in low cost GNSS chips. This work is globally leading-edge, and applications for such performance include navigation, augmented reality and potentially gaming.
⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀
By the end of the summer, Ivan will have furthered the understanding of how applying high-performance measurement processing techniques can improve positioning.

Abstract:

Improving Navigation Satellite-based Positioning Software through Signal Simulations

The objective of this project is to develop accurate satellite range measurement files from a Global Navigation Satellite System (GNSS) signal simulator, that can then be used with York University’s Precise Point Positioning (PPP) software for cm-level positioning. As navigation by GNSS becomes more prominent, the need for better accuracy using single- and dual-frequency receiver chipsets becomes more important. The goal of this research project is to achieve file generation using a GNSS signal simulator that is accurate, repeatable, and highly adjustable. The research initially focused on becoming familiar with all aspects of the GNSS signal simulator and its software package, SimGEN. This work was followed by developing a novel program in MATLAB that creates input files for SimGEN based on the satellite orbits of both GPS and GLONASS satellite systems. The program ensures that the simulated scenarios are as realistic as possible, and also enables the York GNSS Lab to recreate simulated scenarios using actual, field gathered data. The capability to execute accurate simulations and produce output files for York’s PPP algorithm allows for testing of the algorithm. By being able to simulate the measurement errors that are applied, errors can be more accurately modelled, and therefore eliminated. The final aspect of the project is to simulate triple frequency GPS signals. Using three frequencies will increase the accuracy of the user position and reduce solution initialization and convergence to the desired level. The aim of this research is to improve accuracy from tens of metres to the centimetre-level in low cost GNSS chips. This work is globally leading-edge, and applications for such performance include vehicle navigation, augmented reality, and precision agriculture.

 

Brittany Danishevsky, Department of Electrical Engineering & Computer Science
Research Program: NSERC USRA
Supervisor: John Tsotsos

Brittany Danishevsky started at Lassonde this past winter in the Department of Electrical Engineering & Computer Science. Having previously graduated in Psychology, the switch to Computer Science was motivated from first job which exposed her to the potential of technology and AI for positive social impact. Her background in psychology comes in handy for predicting user behaviour and building machine learning algorithms requiring an understanding of attention, vision and perception. 

Brittany's interests meld nicely into her summer research project, where she is working under Dr. John Tsotsos in a computer vision and robotics lab. This summer, she is building software for an autonomous wheelchair with the goal of successfully navigating a cluttered environment using computer vision. This technology can be implemented in nursing homes to relieve some of the burden on nurses and support workers.

Abstract:

Title: Object Detection in Nursing Homes for Autonomous Wheelchair Guidance

When creating autonomous robots for indoor health-care environments, common technologies such as LIDAR, GPS, and RFID chips can be unreliable, and interfere with medical electronics. Thus, computer vision is proposed as a less invasive and more reliable technique. This project, which is divided into two parts, aims to develop software for an autonomous wheelchair, which will use vision to navigate around a nursing home. The first part of this project is building a data set of items encountered in a nursing home (e.g., wheelchairs, walkers, canes, etc.). This dataset will be trained on an existing object detection algorithm (e.g., YOLO9000) and tested in a nursing home setting. The second part of the project is to determine what features of the nursing home can be used to localize the wheelchair, allowing it to autonomously navigate within a pre-mapped space reliably. The strategy for localization is a combination of an existing global pose refinement framework, as well as object classification using the dataset put together in the first part of the project. Successful classification of objects will inform the current location of the wheelchair (i.e., answer “where am I right now?”); global pose refinement, which is guided by wheel odometry and feature detection of wall edges, will allow the wheelchair to localize as it moves through nursing home corridors. Such autonomy will relieve some of the burden on elder care professionals, while also providing independence to nursing home residents. In addition, this project will contribute a novel dataset to the growing body of computer vision research, as well as a unique application of computer vision and robotics to elder care.

 

Mohammadreza Karimi, Department of Electrical Engineering & Computer Science
Research Program: NSERC USRA
Supervisor: Hossein Kassiri

Reza Karimi is a 3rd-year Computer Science student who is working with Dr. Hossein Kassiri this summer. Through using machine learning algorithms to create models that detect seizures, Karimi hopes to have a better understanding of signal processing and optimize the current model in terms of its accuracy and performance by the end of the summer. Although many have worked in this specific field, the creation of commercial products hasn’t been possible due to the many challenges this type of device can face. For example, there’s the issue of limited computational power on small chips. This means that apart from the accuracy and efficiency of the model being made, the product it will be used on must be considered as well. Reza is ambitious to contribute to this project by solving the challenges on his end.

Abstract:

Seizure Detection Using Brain EEG Signal Processing

Approximately 360,000 Canadians live with epilepsy. For 20% of these patients who are refractory to drugs, a medical device capable of early detection of an upcoming seizure could significantly improve their quality of life by either alerting the patient and/or triggering an intervention mechanism such as electrical stimulation.
Over the past decade, recording and processing brain’s electrical activity has been used as a promising method for detecting epilepsy seizures. However, the success has been very limited due to (a) the large patient-to-patient variations in terms of seizure manifestation, and (b) limited computational resources available in an implantable device. Recently, machine learning algorithms have been investigated, with some success, to realize a patient-specific algorithm to overcome the first challenge. However, the majority of reported algorithms are too computationally-expensive for an implantable device.
In this project, we have designed, implemented, and optimized a machine learning algorithm with both detection accuracy and computational efficiency in mind. To achieve high detection sensitivity with minimal false alarm rate, we have extracted various features of the recorded signals such as signal frequency band energy and phase synchronization, and fed them to a trained support vector machine (SVM) classifier known for its reliability and efficiency. Different modules of the implemented algorithm (e.g., data acquisition, feature extraction, and classification) are individually optimized for hardware implementation.
The algorithm is tested on a 916-hour 24-patient labelled MIT EEG Database and can detect 90% of seizures with only a 10 false alarms per day (commercial devices have 200 to 600 daily false alarms (Bergey, et al, Neurology, 2015)).

 

Daphne-Eleni Archonta, Department of Mechanical Engineering
Research Program: NSERC USRA
Supervisors: Pouya Rezai, Khaled Youssef

Daphne-Eleni Archonta is a 2nd-year student in the Department of Mechanical Engineering. This summer, Daphne will be exploring microfluidic devices in the field of medical research in Professor Pouya Rezai’s laboratory. She will be conducting experiments and analysis of the electrotactic behaviour of the Parkinson’s Disease model organism, the C. elegans worm, using microscopy and microdevices. 

Daphne is hoping to further the research on C. elegans’ natural responses and the effect of specific neuronal interactions in the system that control the egg-laying of the worm. This will be achieved with a novel technique utilizing the natural behaviour of the animals for egg extraction and could be used to further the widespread use of C. elegans in Parkinson’s Disease research.

Abstract:

On-demand Electric Field Induced Egg Laying of Caenorhabditis Elegans

C. elegans is a model organism offering a well-mapped and accessible neuronal system for disease studies. Egg-laying is a behavior of interest in neurobehavioral studies, governed by a sensorimotor circuit regulated by different neuronal and muscle cells. Optogenetics, the main method used to stimulate neurons to study the egg-laying sensorimotor pathway, is restricted to genetically-modified worms, calling for a simple, on-demand and inclusive egg-laying stimulation technique applicable to non-mutant worms. 
We report a novel microfluidic technique which exploits electric field (EF) stimulation to induce and investigate egg-laying of C. elegans on-demand, without the need for genetic modification required in optogenetics. A setup was developed to investigate the effect of direct-current EF on stimulating egg-laying in C. elegans partially-trapped in a microfluidic device. The effects of worm age (64-136hr) and EF direction, strength (2-6V/cm) and pulse duration (5-40s) on the number of eggs released within 10min were investigated.
The EF direction and strength were investigated under controlled conditions of the trap, leading to significantly more egg-laying events when 64hr-old worms faced the anode. Moreover, EF exposure time had no significant effect on the total number of eggs laid over 10min by 64hr-old worms. Age had a significant effect on EF-induced egg-laying with 136hr-old worms laying 76% fewer eggs than 64hr-old worms. 
C. elegans EF-induced egg-laying is a novel contribution that provides a simple and on-demand technique to stimulate and investigate egg deposition behaviour and sensorimotor processes in wild-type, and potentially mutant, worms; while not being restricted to genetically-modified strains.

Leave Your Comment

Comments