UMSS22 Engineering and Information Sciences

Abstract: This research presents the development of a novel composite material that will be functional in ocean-going devices and vessels. This thesis presents a desktop model of a novel reinforced concrete base automated manufacturing process. The scale model consists of a two-step process to autonomously create a rebar structure that would be used as reinforcement for a concrete part. In this process, the rebar is first bent into an appropriate shape by an existing automated wire bending machine which is then added to a 3D printed mold. This part creates the mold for the rebar reinforced concrete part. In this process, the mold is printed simultaneously up to the point of rebar insertion. The rebar is then inserted by a robot from the bending machine and printing is continued. These steps are repeated until the mold is completed, upon which the concrete is added. In addition to the bench-scale model, the design and demonstration of full-scale rebar bending is discussed to show feasibility as the process is scaled up to a commercial-scale system. A discussion of the impacts of this new commercial technology is also discussed, as well as its relevant applications. Automated, mass-producible reinforced concrete could revolutionize the world of construction as it is currently known.

Production of Renewable Diesel and Jet Fuel from Biomass-Derived TDO Oil

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Matthew Kline

Sampath A. Karunarathne

Thomas J. Schwartz

Clayton Wheeler

Graduate Student Presentation

Faculty Mentor: Clayton Wheeler

Abstract: As governments around the world place restrictions on the production of energy from fossil fuels, the demand for renewable fuels is projected to increase. One of the most promising materials to convert into bio-oils is cellulose, which is advantageous as a feedstock because residue from logging, farming, or even municipal waste can be utilized. Several processes have been developed to convert cellulose into bio-oil, one of them being Thermal DeOxygenation (TDO), which shows promise. It is a process developed at the University of Maine that converts organic acids from cellulose hydrolysis and dehydration into a low-oxygen bio-oil with a broad boiling point distribution.
While the chemical compounds in TDO oil could be directly used to produce gasoline, they lack some of the characteristics needed for petroleum middle distillates, such as jet fuel and diesel. In order to replace these fuels, TDO oil must be upgraded using catalytic hydrogenation and ring opening reactions.
This poster presentation discusses the chemistry and reactions required to produce renewable gasoline, diesel, and jet fuels for commercial use. Past, present, and future work on upgrading TDO oil will be discussed, as these extra steps are integral in producing an array of transportation fuels. The presentation will also discuss my research on the hydrogenation and ring opening of bicyclic molecules to produce fuels with better diesel combustion characteristics. Using the catalysts presented in this presentation, we can produce diesel and jet fuels from biomass instead of petroleum.

Turtlebot Mapping and Navigation

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Matthew Harzewski

Quinn McCarthy Beaver

Saoirse Douglas

Mason Yu

Vikas Dhiman

Undergraduate Student Presentation

Faculty Mentor: Vikas Dhiman

Abstract: Modern robots are increasingly mobile, needing to observe and navigate an environment autonomously instead of being fixed in place. Devices such as autonomous vehicles, drones, forklifts and even vacuums need to map out their surroundings and avoid obstacles while operating, which necessitates complex combinations of specialized software. Much of this software is experimental in nature, but modular, inhabiting standardized frameworks like Robot Operating System.
The purpose of this project is to evaluate the present landscape of version two of ROS, which uses a new distributed architecture, and construct a functioning stack of software to handle mapping and navigation on the Turtlebot3 hardware. The highly distributed design has presented many challenges, along with a lack of maintained packages and documentation.
Thus far, the core modules have been evaluated and a single robot can successfully map out a room with a laser scanner and drive to given waypoints on the produced map. The core software has also been Dockerized for easier deployment. Additional work is underway to extend this functionality, including work on autonomous exploration to produce maps without human intervention. The latter goal involves writing software to process map data and detect the edge of unexplored space before dispatching the robot to those areas.
The ROS2 ecosystem is promising but presents many challenges in added network issues and software that has not yet been updated. Writing software from scratch is a smooth process due to ROS2 seamlessly taking care of networking and inter process communication.

Combatting Nutrient Scarcity Using Adsorption of Phosphorus from Recirculating Aquaculture System Wastewater

Submission Type: Virtual Presentation

Submission Category: Engineering and Information Sciences

Author(s):

Morgan Oehler

Jean MacRae

Undergraduate Student Presentation

Faculty Mentor: Jean MacRae

Abstract: Recirculating Aquaculture Systems (RAS) treat and recycle process water back into the system, which conserves water and reduces wastewater discharge. RAS wastewater is abundant in phosphorus, a nonrenewable nutrient needed for food production that can cause dead zones when released to surface waters. In this study, a phosphorus adsorbent, RhizoSorb®, was analyzed in RAS wastewater and an artificial seawater solution made up of deionized water, salt, and phosphate to determine if it is well suited for phosphate recovery from RASs. Kinetic and equilibrium adsorption experiments were conducted to compare the adsorption of phosphorus in both solutions and determine the impact of the wastewater constituents on the efficiency of the adsorbent. Wastewater samples were collected at three different salinities (0 ppt, 7 ppt, and 35 ppt) from the United State Department of Agriculture (USDA) Aquaculture Facility in Franklin, Maine. Experiments were conducted on the three different salinities at 10 mgL PO4 concentration for the wastewater samples and artificial wastewater solutions. I found that adsorption capacity decreases with increasing salinity and that the wastewater constituents increase maximum adsorption capacity at high concentrations of salt but had little impact on the adsorption of the sorbent overall. The equilibrium data was better described by the Freundlich isotherm model than the Langmuir isotherm indicating the adsorbent has a variety of sorption sites and is not restricted monolayer coverage. Overall, RhizoSorb ® was effective at phosphorus uptake, and sorption was not greatly impacted by other waste constituents, indicating that is could be used to recover phosphate, especially from freshwater aquaculture waste.

505. The Performance Assessment of a Lighter-than-air Vehicle for Earth Science Remote Sensing Missions

Submission Type: Exhibit

Submission Category: Engineering and Information Sciences

Author(s):

Maxwell Burtis

Wilhelm Friess

Undergraduate Student Presentation

Faculty Mentor: Wilhelm Friess

Abstract: This summer, a lighter-than-air (LTA) drone was tested in Alaska to measure glacier bedrock fracture density and orientation. Five flights were made in low wind conditions, and the directional stability of the airship made it too challenging to control in flight to realistically acquire useful image sets. The directional stability of the airship, when compared to an actively stabilized consumer-grade quadcopter was inferior. Flight logs and GPS data from the GPS on the LTA drone were analyzed and quantitative assessment of the observed instability was made. The yaw axis and pitch were analyzed, and the yaw axis instability was greater than the pitch axis instability. Internal instability was observed as a function of yaw oversteer and the lack of active stabilization. Reduction of the yaw instability was attempted by reducing the yaw motor size and implementing active flight stabilization. The observed instability also resulted from external sources like wind gusts and the glacier microclimate. The analysis informed modifications of the LTA drone to make it more stable for glacier research, which were implemented and tested. Also, recommendation for further modifications include the implementation of autonomous flight control and possibly envelope optimization which may also improve endurance.

Development of a Lightweight Embedded Machine Learning System for Obstacle Avoidance

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Connor Firth

Babak Hejrati

Undergraduate Student Presentation

Faculty Mentor: Babak Hejrati

Abstract: The project we are planning to present for UMSS is focused on the development of a novel system for user-centered autonomous obstacle avoidance. The system consists of a single camera worn by the user, connected to a microcontroller which performs machine learning computations to detect objects in the environment. The novelty of this project lies in the leveraging of recently developed, compact computer vision technology to avoid obstacles in real time on a lightweight system using a new method for obstacle detection. In order to avoid obstacles, we must both detect them within the scene and determine the nearest obstacle to alert the user. Traditional detection methods emphasize identification of possible obstacles, often using object detection algorithms. Our system instead uses image segmentation to detect areas where obstacles do not appear, and analyze the remaining space. This approach avoids problems inherent to the object detection method, such as training dataset limitations and the inability to recognize objects not completely in frame. Additionally, our system is built on a consumer-available foundation, utilizing the Nvidia Jetson Nano as a central processor and several open source Python libraries. This is a significant step from existing human-based obstacle avoidance systems: not only does it use previously cumbersome machine learning networks for monocular depth perception and image segmentation, but it does so in real time on an embedded system. The next steps for this project are the implementation and testing of the system in an experimental setting.

Watch Your Flavors: Augmenting People’s Flavor Perceptions and Associated Emotions Based on Videos Watched while Eating

Submission Type: Virtual Presentation

Submission Category: Engineering and Information Sciences

Author(s):

Meetha Nesam James

Nimesha Ranasinghe

Graduate Student Presentation

Faculty Mentor: Nimesha Ranasinghe

Abstract: People engage in different activities while eating alone, such as watching television or scrolling through social media on their phones. However, the impacts of these visual contents on human cognitive processes, particularly related to flavor perception and its attributes, are still not thoroughly explored. This paper presents a user study to evaluate the influence of six different types of video content (including nature, cooking, and a new food video genre known as mukbang) on people’s flavor perceptions in terms of taste sensations, liking, and emotions while eating plain white rice. Our findings revealed that the participants’ flavor perceptions are augmented based on different video content, indicating significant differences in their perceived taste sensations (e.g., increased perception of salty and spicy sensations). Furthermore, potential future implications are revealed to promote digital commensality and healthier eating habits.

Effects of Harsh Environmental Conditions on Mechanical and Physical Behaviors of Tough and Stretchable Hydrogels for the Benefit to Optical Fiber Technology

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Stella Cotner

Siamak Shams Es-haghi

Undergraduate Student Presentation

Faculty Mentor: Siamak Shams Es-haghi

Abstract: Aboard spacecrafts, use of optical fibers has been limited by their mechanical properties. Optical fibers are made from drawn glass or plastic, which becomes brittle and prone to fracture in the harsh climate of outer space. If optical fibers could remain flexible and durable in the climate of space, the communication and data transfer rate in spacecrafts could be greatly improved. Hydrogels are networks made from long molecular chains that have been swelled in aqueous solution and have unique properties. Loosely crosslinked double-network (DN) hydrogels were synthesized and swelled in an aqueous glycerol solution and a sodium chloride solution in water at specific concentrations in order to make a tough hydrogel that is freeze-tolerant. Samples of glycerol- and sodium chloride- swelled hydrogels were initially tested along with water-swelled control samples using Differential Scanning Calorimetry (DSC) to determine the thermal transitions of the materials. The hydrogel samples underwent freeze/thaw cycles and their mechanical behavior under loading-unloading-reloading large-strain tensile deformation were analyzed. The results of these tests indicate that all the data points for deformation paths coincide showing no mechanical damage happens during deformation. Similar results were observed for samples boiled in their swelling solutions. The results of this study show that the DN hydrogels can sustain harsh environmental conditions and therefore are potential candidates for space-related applications.

CEU-Net: Ensemble Semantic Segmentation of Hyperspectral Images Using Clustering

Submission Type: Virtual Presentation

Submission Category: Engineering and Information Sciences

Author(s):

Nicholas Soucy

Salimeh Yasaei Sekeh

Graduate Student Presentation

Faculty Mentor: Salimeh Yasaei Sekeh

Abstract: Most semantic segmentation approaches of Hyperspectral images (HSIs) use and require preprocessing steps in the form of patching to accurately classify diversified land cover in remotely sensed images. These approaches use patching to incorporate the rich neighborhood information in images and exploit the simplicity and segmentability of the most common HSI datasets. In contrast, most landmasses in the world consist of overlapping and diffused classes, making neighborhood information weaker than what is seen in common HSI datasets. To combat this issue and generalize the segmentation models to more complex and diverse HSI datasets, in this work, we propose our novel flagship model: Clustering Ensemble U-Net (CEU-Net). CEU-Net uses the ensemble method to combine spectral information extracted from convolutional neural network (CNN) training on a cluster of landscape pixels. Our CEU-Net model outperforms existing state-of-the-art HSI semantic segmentation methods and gets competitive performance with and without patching when compared to baseline models. We highlight CEU-Net’s high performance across Botswana, KSC, and Salinas datasets compared to HybridSN and AeroRIT methods.

Design, Implementation, and Test of Spacecraft Antennae and a Ground Station for MESAT1

Submission Type: Virtual Presentation

Submission Category: Engineering and Information Sciences

Author(s):

Travis Russell

Joseph Patton

Steele Muchemore-Allen

Ali Abedi

Graduate Student Presentation

Faculty Mentor: Ali Abedi

Abstract: MESAT1 is a CubeSat that was proposed by the University of Maine in response to NASA’s CubeSat Launch Initiative, and in early 2020 was selected by NASA to be launched into a Low Earth Orbit (LEO) in June of 2022. The satellite will carry four low cost complementary metal–oxide–semiconductor (CMOS) cameras which serve as sensing instruments for three science missions proposed by K-12 schools in Maine. The cameras periodically take pictures of Earth to analyze water turbidity, identify urban heat islands, and predict harmful algal blooms. The multi-spectral image data is packed into frames and downlinked as BPSK digital data at a rate of 1200 bits per second using a 100 mW Ultra High Frequency (UHF) transmitter. A ground station is required to receive the data and control the satellite via uplink commands using a Very High Frequency (VHF) radio. A UHF and VHF antenna are required to interface with the spacecraft radio and allow for transmission and reception of signals. To fulfill these requirements, a ground station was built on the University of Maine Orono campus that allows for communication with the MESAT1 spacecraft at low elevation angles. Two quarter-wavelength monopoles were designed with relatively isotropic radiation patterns and low mismatch losses at the command uplink (VHF) and telemetry frequencies (UHF).

Roll-to-Roll Manufactured Custom Textured Films for Low-Cost Pulsed Electric Field Water Purification Devices

Submission Type: Virtual Presentation

Submission Category: Engineering and Information Sciences

Author(s):

Liza White

Amelia Cobb

Kyaira Grondin

Caitlin Howell

Graduate Student Presentation

Faculty Mentor: Caitlin Howell

Abstract: Water purification and disinfection, particularly of turbid water, is a significant and growing problem worldwide. Recent research at the Wyss Institute has resulted in a device that uses a pulsed electric field (PEF) to inactivate pathogens in water; however, manufacturability, power consumption, and cost remain significant hurdles. Through leveraging paper industry technology in Maine, we looked to optimize electric field generation using a custom textured film in a roll-to-roll manufacturing process. Specifically, we used Maine paper companies’ release paper technology as a substrate and explored different types of sputtered metal and coatings to reduce the overall power consumption and cost as well as improve the manufacturability. Mass-manufactured textured materials were cut into singular flow cells, sputter-coated with various metals, and assembled. The flow cells were connected to a pulsed generator that pulsed a square wave at 15 µs at a frequency of 100 Hz with voltages ranging from 50-200 V. Through these tests, we found that the sputter-coated textured material was able to conduct electricity effectively. Additionally, CAD and modeling software simulated various textures to determine the optimal texture to focus the electric field while keeping a low total current density. The simulation data showed that a sawtooth pattern with a slight rounded peak performed the best with a maximum electric field intensity of 3.67 V/µm and a current density of 0.002 mA/µm2. These results demonstrate the feasibility of paper industry roll-to-roll manufacturing processes to produce materials in low-cost PEF water purification devices.

JMG Student Site Check-In Application – Presentation withdrawn from judging

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Elijah Caret

Spencer Morse

Michael Ferris

Xingzhou Luo

Terry Yoo

Undergraduate Student Presentation

Faculty Mentor: Terry Yoo

Abstract: Jobs for Maine Graduates (JMG) is a non-profit organization that specializes in helping students, from middle to post secondary education, in career oriented subjects such as applying for jobs, colleges, and internships. They also have an extended learning opportunities (ELO) program. ELOs are out-of-school events/sessions/internships that students can take part in for school credit. We were approached by ELO Director Samantha Brink and ELO Program Manager Lanet Anthony regarding new software. They are looking for another way to track the necessary data for this program because the current CRM system and education platform that they use don’t fit that need. Our solution is a web application that will allow students to “check-in” to their particular ELO, this will then be logged on the application. Supervisors at these ELOs will then be able to approve of these check-ins and grade the student’s performance. The JMG representative teacher at that student’s school will be able to see this information as well. We are currently developing the application using a platform known as budibase, which will help dramatically reduce the amount of code needed to write. Our goal is to deliver an application that is easy to use for students and easy to maintain for the JMG staff by the end of the semester. We are adopting both an AGILE and process-driven software development methodologies involving heavy documentation and planning with iterative development. We are currently in the development phase of the project.

The University of Maine Digital Program of Study Approval System – Presentation withdrawn from judging

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Mackenzie Creamer

Liam Blair

Aaron Wilde

Vincent King

Peter Riehl

Terry Yoo

Undergraduate Student Presentation

Faculty Mentor: Terry Yoo

Abstract: Our project, The Graduate School Program of Study Appoval System (POSAS), serves as a transition from the University of Maine’s physical storage of Programs of Study (POS) documents, to the digital storage of these documents. Currently, POS documents for graduate students are held in physical documents stored at the University of Maine. These physical documents are cumbersome to maintain and update, as advisors must approve even small changes made to these documents. Specifically, the SCIS graduate programs here at the University of Maine have grown larger, and the current problems from the physical storage of these documents can only become more time consuming and problematic as the department grows larger. Thus, the goal of this project is to completely digitalize future POS documents through a University of Maine hosted server that allows graduate students and advisors to work through editing POS documents as well as sign them electronically. Digitalization of the POS documents will greatly reduce the time invested and effort involved with editing and signing these POS documents for graduate students and advisors.

While not done yet, we are aiming to have a functional project by the end of the semester as a proof of concept for the larger graduate school as a whole.

Scientific Terminology Extractor (SciTEx)

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Gunnar Eastman

Torsten Hahmann

Undergraduate Student Presentation

Faculty Mentor: Torsten Hahmann

Abstract: While there are many advances being made in researching and producing new types of cellulose-based materials, the research publications announcing them use highly specialized and sometimes inconsistent technical language, which can make it difficult to know what exactly is new about the materials being produced and how each of them could be used. Thus, it is important to have a means by which to standardize the terminology used to describe such materials to make it easier to search and explore the scientific literature.
The goal of this research is to develop software that can identify the specialized terminology in need of standardization as an important tool towards improving access to the knowledge in scientific publications. The program will take PDF documents as input and read them, while maintaining document structure and keeping track of citations. It will then extract words that are neither names nor in a standard dictionary, and return this list of technical terms and phrases, as well as the location(s) of each term in the document. The tool will be evaluated by collecting scientific articles and separating them into two groups: one will be used for testing during development, the other will be used to measure the program’s performance.
The tool will be of broader use to all scientific fields, as it could either return a list of all technical terms, beyond those related to cellulose-based materials, or it could be further calibrated to produce a more refined list tailored to specific fields.

Top-Down Approach to Super Flexible and Antibacterial Wood

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Meghan Boos

Caitlin Howell

Yingchao Yang

Undergraduate Student Presentation

Faculty Mentor: Yingchao Yang

Abstract: Airborne microbes are a major cause of respiratory ailments, such as allergies and pathogenic infections. A soft material having capability to inhibit microbe growth can be integrated into face masks to address the spread of microbes. Silver nanoparticles (AgNPs) have been used effectively in previous research to prevent the growth of a range of clinically relevant bacteria. However, the inhalation of AgNPs can be harmful to the human body. In this research, balsa wood was chemically treated to have excellent flexibility and pores as a substrate to house AgNPs, demonstrated with the material’s ability to inhibit bacterial growth. The hybrid biomaterials were characterized by Scanning Electron Microscopy, Energy Dispersive X-Ray Analysis, and bacterial zone of inhibition tests to determine the morphology, chemical deposition, and antibacterial properties, respectively. The results showed that the untreated balsa wood has irregular hexagonally shaped pores and the chemically treated balsa wood had smaller, more crumpled pore shapes due to the partial removal of lignin in wood which allows for greater deformation of pores, therefore making the material super flexible. Using hydrothermal methods, AgNPs were grown in the pores of the chemically treated balsa wood. AgNP clusters of approximately 2.5-5 micrometers were successfully synthesized. Zone of inhibition tests on 0.02M AgNO3, and control samples were used to analyze the growth-inhibition capabilities of AgNPs on Escherichia coli. Our results demonstrate the creation of super flexible and antibacterial wood material to help solve the issue of bacterial resistance on flexible medical devices for people around the world.

Key Characteristics Analysis of Stanford Drone Dataset

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Josh Andle

Salimeh Yasaei Sekeh

Graduate Student Presentation

Faculty Mentor: Salimeh Yasaei Sekeh

Abstract: The Stanford Drone Dataset (SDD) is a widely used trajectory dataset, containing annotated information of individuals navigating various scenes. Such datasets are useful for many real-world tasks, including trajectory prediction and object tracking. Despite its prominent use, discussion surrounding the characteristics of this dataset is insufficient. In this work, we aim to demonstrate several characteristics of the SDD which aren’t readily reported elsewhere, and how not properly accounting for these characteristics can impact model performance. Our analysis of the SDD’s key characteristics is important because without adequate information about available datasets a user’s ability to select the most suitable dataset for their methods, to reproduce one another’s results, and to interpret their own results are hindered.
In order to visualize these characteristics we utilize an information-theoretic measure and custom Adaptive Interaction Measure (AIM). We additionally confirm these findings when possible by demonstrating the impact of these characteristics on a previously published trajectory prediction model. Lastly, we compare and contrast the SDD with the similar Intersection Drone (inD) dataset to study their characteristics, and to emphasize the importance of selecting the most appropriate dataset for a given application. The observations we make through this analysis provide a readily accessible and interpretable source of information which will help improve performance and reproducibility of methods applied to this dataset going forward.

Smart Cane for the Visually Impaired

Submission Type: Exhibit

Submission Category: Engineering and Information Sciences

Author(s):

Jordyn Judkins

Sarah Clark

Danika Evangelista

Caden Scott

Undergraduate Student Presentation

Faculty Mentor: Robert Bowie

Abstract: The visually impaired often use blind canes in order to navigate independently. This project aims to assist a visually impaired user by detecting obstacles at range and providing non-visual feedback to the user in order to allow them to maneuver independently. Previous designs for these devices, termed smart canes, have focused on ultrasound techniques for detection while providing feedback through vibrations and sounds; however, they cause unnecessary distractions to the user in disorganized vibrations or extraneous noise. To do this, we used a series of sensors that provides the user with organized tactile feedback based on a range of detection. A ball-point pen style tip was used to allow for some impact absorption which lessens the stress on the user’s hand and arm providing more comfort. The design was tested for time to change course, available speed of travel, number of objects hit, and notification time. The tests showed that the design was able to notify the users of obstacles in an acceptable amount of time for them to change course before hitting an object. This showed that the newly designed smart cane provides better feedback and increased comfort for the visually impaired user.

Acknowledgements:
The authors would like to thank Angel Hildreth, Keith Hodgins, Dr. Robert Bowie, Liza White, and Dr. James Beaupre for materials, brainstorming, critical advice, and guidance in this project.

The Influence of Heat and Mass Transfer on the Setting Rate of Adhesives Between Porous Substrates

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Mubarak Khlewee

Douglas Bousfield

Graduate Student Presentation

Faculty Mentor: Douglas Bousfield

Abstract: With the move to replace plastic packaging with sustainable options, glueing operations are needed to form cellulose based packaging. In the setting of hot melt adhesives and water-based glues in the production of paper-based packaging, the controlled penetration of the adhesive is important to obtain rapid setting rates and good bond strength. The penetration of the polymer into the pore space influences the product performance and the ability of the product to be recycled at the end of life. Also, the dynamic flow of fluid into a porous media considering temperature or concentration changes is of interest for a number of other industrial processes. Experiments are designed to understand the extent of penetration of hot melt and water-based adhesives into several porous substrates. Paper surfaces are modified by a range of coatings that have different porosities and pore sizes, and contact angles; these surfaces were characterized with a range of techniques. For hot melt adhesives, a layer of the adhesive is pressed against the paper of interest with the Carver press, for a known time, pressure, and temperature. The final degree of penetration of both systems is determined with silicon oil, mercury intrusion porosimetry, thickness, and weight methods. Here, the accuracy and repeatability of these methods are compared. The strength of the bond at various setting conditions is also measured with the mechanical tester (INSTRON). Using two different press temperatures and pressures, the four methods gave excellent matching results regarding adhesive penetration depths. The mechanical tester (INSTRON) results showed that for a low press temperature, the failure was an adhesive failure mode.
While there is much experimental work discussing various parameters in these operations, little theoretical work has been reported. In this regard, various models are developed to predict the penetration of the adhesive as a function of the fundamental parameters. A finite element method-based model (COMSOL Multiphysics) is used to solve the unsteady-state flow of liquid adhesives into a single pore or a porous medium accounting for temperature or concentration changes, where the fluid viscosity is a function of temperature or concentration. The model predictions are compared to experimental results. Good agreement for different paper types, pressing times, and temperatures are obtained. For the hot melt adhesive, the model indicates that the substrate temperature is a critical parameter in the penetration of fluid because the viscosity of the fluid can increase a large amount due to cooling in the pore. For the water-based adhesive, the model indicates that when the solvent diffuses through the pore walls, the adhesive particles will accumulate near the pore walls, leading to an increase in its concentration. This increase in concentration will lead to an increase in the viscosity of the adhesive, decreasing the rate of adhesive penetration.

Taste and Odor Degradation in Water by Nanobubble-Facilitated Ultrasonication

Submission Type: Exhibit

Submission Category: Engineering and Information Sciences

Author(s):

Zachary Doherty

Sudheera Yaparatne

Onur Apul

Undergraduate Student Presentation

Faculty Mentor: Onur Apul

Abstract: Treatment of tainted water for drinking water applications is a constant battle, often fought against the presence of earthy or musty odors and tastes. These odors or tastes are a result of the increased production of volatile organic compounds by certain species of algae. Most often the odorous algal metabolite compounds i.e., geosmin (GSM) and 2-methylisoborneol (MIB) are identified as the culprit behind taste and odor issues in drinking water. While many factors such as temperature, microbial concentrations, media, and biodegradable organic matter can affect the removal of these compounds there yet lacks a singular sustainable strategy to degrade the GSM and MIB compounds. This project will explore the use of cutting-edge nanobubble technology for the removal of GSM and MIB. The primary focus of this research is the interaction between intentionally ruptured nanobubbles and the impact it has on GSM and MIB compounds. Ultrasonication is used to induce the nanobubble collapse and thereby generation of hydroxyl radicals is achieved for the degradation of MIB and GSM. Unique behavior and the immense potential of nanobubble technology has been observed for wastewater treatment applications in our earlier studies by 5-6 folds faster mass transfer and up to 50% more chemical removal and the current project allows further opportunities to explore new aspects of nanobubble impact on water treatment.

A Critique of America’s Public Transportation – Presentation withdrawn from judging

Submission Type: Virtual Presentation

Submission Category: Engineering and Information Sciences

Author(s):

Nathaniel French

Meredith Kirkmann

Undergraduate Student Presentation

Faculty Mentor: Meredith Kirkmann

Abstract: The state of public transportation in the United States is inadequate, ineffective, and needs investment in comparison to many other countries. Throughout the Spring 2022 semester, I am researching the culture and state of public transportation in the United States. The current state of public transportation in the United States includes subways, trains, bus systems, and ferries. The public transportation system is often inefficient or ineffective compared with other similar economies. An in-depth exploration into the public transportation in Portland, Maine will cover the current public transportation system, its deficits, and compare it to the public transportation system of a similar-sized international city that is more effective and adequate for the population. This research includes a review of the Greater Portland Council of Government’s (GPCOG) public transportation planning documents, bicycles, and pedestrian plans, and current analysis of commuting in the Greater Portland area, as well as an interview with Andrew Clark, Transportation Planner at GPCOG. Current legislation and funding are intended to improve the public transportation system in the United States, however, planning for and leveraging that funding to best improve Portland, Maine, and the United States public transportation system is a complex problem. Lessons from other countries can help to inform this work. Connections will be made between suggested future improvements to our public transportation system based on more efficient and effective systems worldwide.

Smart Wheelchair Seat Cushion

Submission Type: Virtual Presentation

Submission Category: Engineering and Information Sciences

Author(s):

Kacey Roehrich

Joshua Leake

Chisom Orakwue

Madison Raza

Lisa Weeks

Undergraduate Student Presentation

Faculty Mentor: Lisa Weeks

Abstract: Pressure ulcers (PUs) are a common and often avoidable skin and soft tissue injury that afflicts wheelchair users. Current wheelchair cushions designed to minimize PUs provide support and evenly distribute pressure across the ischial tuberosities (ITs); however, these cushions do not address the need for users to routinely shift their weight and restore blood flow to the soft tissues of the gluteal region. To address the need for reducing PUs in wheelchair users, we designed and built a wheelchair seat cushion that can dynamically vary the pressure under the patient to automatically redistribute weight and reduce periods of occluded circulation. A prototype cushion utilizing blood pressure cuffs as air bladders, diagram pumps, solenoid valves, and pressure transducers was developed and tested with the system interfaced through a Raspberry Pi. Testing procedures quantified the accuracy and precision of the pressure transducers against standard gauge pressure, pump efficiency in its ability to supply air under load pressure, and a pressure mapping model which modeled the physics of the cushion and contact forces on the ITs. Preliminary results showed the pressure transducers were within non-statistically significant error when compared to a pressure gauge, the pumps supplied air and held pressure under compression, and the cushion maintained a physiologically appropriate pressure on the ITs with 20-minute cycling over four ranges of pressures across the blood pressure cuffs. Future tests should involve using a pressure mapping system to quantify the efficacy of the cushion in reducing pressure-related injuries. The results demonstrate that our smart wheelchair seat cushion can help reduce PUs by cyclically varying the pressure within the cushion, thus reducing the pressure exerted on the ITs, which will solve the problem of pressure-related injuries in wheelchair users around the world.

Low-Cost Electrical Power System for CubeSats

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Joseph Patton

Steele Muchemore-Allen

Travis Russell

Ali Abedi

Graduate Student Presentation

Faculty Mentor: Ali Abedi

Abstract: An electrical power system (EPS) is a system of a spacecraft that is responsible for power generation, storage, distribution, and control. An EPS must be efficient, reliable, and it must distribute power safely to all other spacecraft systems. The MESAT1 EPS is a low-cost custom-designed EPS circuit for the MESAT1 CubeSat. A commercial off-the-shelf (COTS) EPS can cost tens of thousands of dollars, while the MESAT1 EPS can be produced for a cost of about $500. Because COTS EPS units must be designed to meet the power needs of a wide variety of spacecrafts, they may be built to handle a wide range of loads at the cost of efficiency. The MESAT1 EPS is designed to be highly efficient at the specific load requirements of the MESAT1 spacecraft, which results in an overall power savings VS a COTS EPS. While remaining at a very low cost, the MESAT1 EPS features important safety and reliability elements such as battery balancing and temperature control, overcurrent and undervoltage protection, and automatic reporting of EPS system status, and efficiency boosting features such as maximum peak power tracking. In this study we discuss the design and testing of the MESAT1 EPS.

Privacy-Based Access Control Model for Blockchain-stored Healthcare Data

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Dilrukshi Abeyrathne

Silvia Nittel

Graduate Student Presentation

Faculty Mentor: Silvia Nittel

Abstract: Healthcare data often contain sensitive personal information, which raises various privacy concerns. Additionally, these sensitive data are currently dispersed across several information systems owned by different healthcare organizations, thereby aggravating this matter. At present, patients have limited access and little control over who has access to which part of their healthcare data; individual patients’ privacy preferences are not supported. For instance, an HIV patient may prefer to have a higher level of restriction in accessing their healthcare records than a patient with a diabetic condition. In recent years, research studies have proposed using blockchain technology to support the transfer of ownership of medical records to the patient to protect their privacy. However, blockchain technology alone is not adequate to provide fine-grained access control and ensure privacy. In this work, we develop MyHealthChain, a blockchain-based healthcare record system that includes a novel record structure that supports medical records and preserves privacy. As our main contribution, we introduce and implement the MyHealthChain’s Privacy-Based Access Control model (PBAC) to help regulate access to blockchain-stored healthcare data based on a privacy attribute that we define as the Privacy Class. A Privacy Class value is computed based on a function that takes a patient’s preferences and the record into account; it is computed per medical record. The record’s owner, the patient, can customize a privacy level through this novel function. Thus, instead of having few privacy levels that are the same for all the patients, the proposed PBAC model facilitates the patient to tailor the privacy based on their needs and preferences.

Techno-economics and Life Cycle Analysis of Upgrading Woody Biomass to Diesel Blendstock

Submission Type: Virtual Presentation

Submission Category: Engineering and Information Sciences

Author(s):

Aysan Najd Mazhar

Thomas Schwartz

Sampath Gunukula

Clayton Wheeler

Graduate Student Presentation

Faculty Mentor: M. Clayton Wheeler

Abstract: Woody biomass conversion to transportation fuels have been developed as alternatives to fossil fuel production to reduce greenhouse gas emissions and to increase energy security. Fast pyrolysis, a thermochemical technology, has the potential to offer high efficiencies to produce liquid transportation fuels from woody biomass. Fast pyrolysis involves rapid heating of biomass particles in the absence of air at approximately 500°C and results in non-condensable gases, bio-oil (pyrolysis oil), and char as products. The pyrolysis oil can be upgraded via integrated mild hydrogenation and etherification processes to a renewable diesel blendstock. Predictive models indicate that many ethers, which might be produced from components in pyrolysis oils, would have good blendstock characteristics, such as enhanced cetane and reduced sooting, at the expense of an energy density that is lower than petroleum diesel.
A techno-economic analysis (TEA) and a life cycle analysis (LCA) have been conducted for the integrated process. A simulation was created using Aspen Plus®, and data for fast pyrolysis at 500 °C was used for the base case. Thermodynamic properties of hydrogenation and etherification model compounds were estimated using Density Functional Theory (DFT). The results from the process simulation were used to estimate the capital and operating costs for a plant that processes 2,000 dry metric tons per day of forest residues. The well-to-wheels LCA was developed using Argonne’s GREET 2019 software to assess the environmental sustainability of producing the renewable diesel fuel blendstock.
The TEA determined that diesel blendstock products with a lower heating value of 23.33 MJ/kg might be produced from forest residues at a minimum fuel selling price (MFSP) of $2.99/ Diesel Gallon Equivalent (DGE).The requirement to remove a lower amount of oxygen from the pyrolysis oil, compared to complete hydrodeoxygenation, means that reforming of non-condensable gases from the pyrolysis reactor provides sufficient hydrogen for the mild hydrotreating reactor. Temperature effects on the pyrolysis oil composition and yields predict that the MFSP might be reduced to $2.74 per DGE at lower pyrolysis temperatures. The well-to-wheels LCA analysis indicates that the renewable diesel blendstock would have a 93% reduction in fossil fuel use and a 93% reduction in greenhouse gas emissions relative to petroleum diesel.

Effect of Interface Contact Conditions on the Electrical Resistance of 3D-Printed Conductive Filaments

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Mackenzie Ladd

Brett Ellis

Undergraduate Student Presentation

Faculty Mentor: Brett Ellis

Abstract: A recent innovation in the world of 3D printing has been the introduction of electrically conductive filaments for household 3D printers. Such filaments allow for a broad spectrum of novel 3D-printed electro-mechanical devices, such as simple circuits, sensors, motors, and switches. Although promising, this technology is limited by inconsistent electrical resistance characterization methods and a lack of reliable material characterization data. This research addresses these problems by comparing resistance characterization methods (e.g., alligator clips, embedded copper wire) published in the literature, quantifying electrical resistance as a function of the probe force, numerically determine contact surface area as a function of probe force and geometry, and comparing results to those found in literature. Electrical resistance as a function of probe force was determined on Protopasta Conductive PLA 20 mm × 20 mm × 20 mm specimens. Specimens were printed at 220°C with a 60°C bed temperature, at a 0.125 mm layer thickness and with 100% 45°-45° opposed rectilinear infill. Results indicate electrical resistances published in the literature fail to account for: (1) resistance at the probe-filament interface, and (2) influence of probe tip force. Such results are critical to accurately determine the effects of process and micro- and meso-structure on electrical resistance. Emphasis will be placed on the methodology to collect precise and repeatable resistance data of conductive materials. This research is important in that it is foundational to the study of electro-mechanical systems and process- and structure-dependencies of conductive filaments.

Geospatial Data Portal – Presentation withdrawn from judging

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

David Sincyr III

Anthony A. Jackson

Grant Shotwell

Stephen Kaplan

Devin Carter

Terry Yoo

Undergraduate Student Presentation

Faculty Mentor: Terry Yoo

Abstract: Geospatial data—information describing phenomena and objects located on or near Earth’s surface—is all around us, and it increasingly powers the infrastructure and services that drive our internet-connected society. Large-scale analysis of geospatial data is helping researchers and commercial enterprises alike discover new or more efficient ways to tackle the world’s biggest problems, from creating optimized package delivery systems to designing infrastructure resilient to climate change. Solving these monumental problems involves an equally monumental amount of geospatial data and processing power to match, and scalable cloud-based analysis platforms are particularly well-suited to meet these resource needs. To match the growing needs of researchers and businesses, our team is working to design and implement the Geospatial Data Portal, a web-based system for managing, viewing, and sharing geospatial data projects. The Portal aims to provide users with a cloud-hosted platform allowing them to upload geospatial datasets, create and execute Python scripts for manipulating geospatial data and generating insightful data visualizations, and share their analyses and visualizations with others. In addition, our users will be able to create groups to work on projects together or make their data publicly available so that anyone can use it in their projects. These functionalities will be made possible through integrating Amazon’s highly scalable, reliable, and robust infrastructure for storing and managing data. Amazon’s services will also provide significant security for the project: utilizing Amazon Lambda to securely run user-supplied code without threatening the server with malicious attacks. To meet the needs of users with ambitious and diverse goals, our team aims to maximize the utility and scalability of the system; that is, we have designed the system to provide geospatial project managers the tools and flexibility necessary to work with a wide variety of data types, file types, and goals, and to accommodate their resource needs at short notice without breaking their data analysis script or noticeably disrupting their workflow.

Identifying the Magnitude and Character of Microplastic Pollution in Frenchman Bay, Maine

Submission Type: Exhibit

Submission Category: Engineering and Information Sciences

Author(s):

Jessica Liggiero

Taylor Bailey

Dilara Hatinoglu

Lauren Ross

Onur Apul

Undergraduate Student Presentation

Faculty Mentor: Onur Apul

Abstract: With the rising global crisis of microplastic contamination in aquatic systems, this study aims to determine the source of microplastics pollution in Frenchman Bay in Maine for the first time. Our work aims to answer the following research questions: Do microplastics come from the freshwater streams, surface runoff, and rivers entering the bay or from the vast open ocean or are they locally generated by fishing activities? Within two seasons over four months, 99 water samples were collected from Frenchman Bay from the coast. There were eight different sites, where five were saltwater and three were nearby freshwater feeding into the Bay. In addition, this research investigated correlations between the presence of microplastics and the turbidity of the water to understand if microplastics behave like other natural colloids. In terms of methods used, at each site 30 liters of water were filtered through a fine-mesh sieve and samples were cataloged in the lab using a microscope. Results indicate that there were denser colloids in the freshwater samples, but more microplastics found in the saltwater samples with less dense sediment. From this research, we found some preliminary conclusions that show the microplastics come from the ocean or localized fishing activities. We have found a link between low turbidity or sediment and elevated microplastics.

The Exploration a Power Spectrum Analysis in Comparison with the Two-Dimensional Modulus Maxima (2D WTMM) in Fractional Brownian Motion (fBm) Images

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Basel White

Andre Khalil

Undergraduate Student Presentation

Faculty Mentor: Andre Khalil

Abstract: The CompuMAINE Laboratory utilizes a powerful multifractal method called Two-Dimensional Wavelet Transform Modulus Maxima (2D WTMM) method to develop predictive diagnostic tools in biomedicine. To study tumor-prone breast tissue from mammograms, a sliding window approach based on the 2D WTMM method has been successfully developed, based on the extraction of the Hurst (H) roughness exponent. However, the approach is attributed to increased computational difficulty and time. With this, the goal of this research is to explore replacing or combing the 2D WTMM approach to the calculation of H with a power spectrum analysis proven to be more computationally efficient, albeit less versatile. Our hypothesis is that the power spectrum analysis is more accurate than the 2D WTMM approach with monofractal images associated with H values less than 0.5, and the opposite for H values greater than 0.5. In order to achieve verify this hypothesis, a roughness analysis is being conducted on fractional Brownian motion (fBm) images through the use of both methodologies in which several calibration parameters and the wavelet of choice for the 2D WTMM methodology are being studied to observe changes in computational efficiency. The utilization of the Mexican hat wavelet function, rather than the Gaussian function, has been shown to increase computational accuracy for fBm images H values greater than 0.5. The variation of coefficient of determination (R2) values in both the 2D WTMM and power spectrum analysis is currently under investigation. With more results from the investigation of these two analyses, future research entails the application of the optimized methodology to the gridding roughness analysis on grayscale mammograms.

Metal 3D Printed Conformal Cooling Channels

Submission Type: Exhibit

Submission Category: Engineering and Information Sciences

Author(s):

Liam King

Branden Hebert

David Winchell

Connor Larson

Brett Ellis

Undergraduate Student Presentation

Faculty Mentor: Brett Ellis

Abstract: Recently enabled via powder-based additive manufacturing of metals (PBAMM), conformal cooled injection molds have reduced cycle times by up to 12%, and decreased warpage and residual stresses of molded parts by enabling tailorable cross-sectioned, non-prismatic cooling channels to be placed adjacent to molding surfaces. Although useful, PBAMM molds are expensive. This research seeks to address this problem by exploring the suitability of Bound Metal Deposition additive manufacturing of metals (BMDAMM), costing ~70% less than PBAMM, for injection mold applications. Experimentation included the design, manufacture, and testing of a 76.2 mm × 50.8 mm × 15.2 mm 316L benchmark component. The component contains two 1/8” BSPP water fittings, thermocouple and heating element ports, and an irregular-pentagon cross-sectioned cooling channel having an S-shaped path through the infilled component. The infill is oriented in the build direction, thus imbuing transversely isotropic properties at the mesoscale. After heating to a temperature of approximately 300°F, heat transfer properties will be characterized via thermal imaging of the surface, thermocouple measurements within the component, and mass flow and temperature measurements of the approximately 50°F cooling water. Next, the cooling channel will be sealed and pressurized to 10,000 psi to determine the pressure at which the pressure seal of the cooling channel is lost. Finally, the component will be compressively loaded to simulate pressure exerted by the molten material during injection. Results from the three tests will be compared to values estimated via finite element analysis thus indicating the suitability of BMDAMM injection molds.

Deformable ProtoPNet: An Interpretable Image Classifier Using Deformable Prototypes

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Jon Donnelly

Alina Jade Barnett

Chaofan Chen

Undergraduate Student Presentation

Faculty Mentor: Chaofan Chen

Abstract: Machine learning has been widely adopted for many high-stakes applications such as healthcare, finance, and criminal justice. To address concerns of fairness, accountability and transparency, predictions made by machine learning models in these domains must be interpretable. One line of work approaches this challenge by integrating case-based reasoning with deep neural networks to produce accurate yet interpretable image classification models. These models generally classify input images by comparing them with prototypes learned during training, yielding explanations in the form of “this looks like that.” However, methods from this line of work use spatially rigid prototypes, which cannot explicitly account for pose variations. In this paper, we address this shortcoming by proposing a case-based interpretable neural network that provides spatially flexible prototypes, called a Deformable ProtoPNet. In a Deformable ProtoPNet, each prototype is made up of several prototypical parts that adaptively change their relative spatial positions depending on the input image. This enables each prototype to detect object features with a higher tolerance to spatial transformations; consequently, a Deformable ProtoPNet can explicitly capture pose variations, improving both model accuracy and the richness of explanations provided, thus enabling wider use of interpretable models for computer vision.

Low-Cost Forest Senor Networks Reliability and Wireless Capabilities

Submission Type: Virtual Presentation

Submission Category: Engineering and Information Sciences

Author(s):

Thayer Whitney

Sonia Naderi

Ali Abedi

Undergraduate Student Presentation

Faculty Mentor: Ali Abedi

Multiscale Anisotropy Analysis of Second-Harmonic Generation Imaging of Pancreatic Cancer

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Joshua Hamilton

Karissa Tilbury

Peter Brooks

Andre Khalil

Graduate Student Presentation

Faculty Mentor: Andre Khalil

Abstract: Pancreatic ductal adenocarcinoma (PDAC) is one of the most lethal cancers known to medicine with a median survival time of 6 months from diagnosis. PDAC is typically characterized by a dense, stiff tumor microenvironment from hyperactive synthesis of collagen resulting in fibrosis. This fibrosis is protumorigenic and limits the penetration of existing therapies. Interestingly, through animal studies, multiple attempts to reduce collagen deposition resulted in more lethal and aggressive tumors with less connective tissue. A new tool is needed to quantitatively characterize the underlying structural changes in collagen, to increase understanding of the dynamics of the tumor microenvironment. Second Harmonic Generation (SHG) provides a label-free, collagen-specific imaging method with submicron resolution. Forty PDAC H&E slides were annotated by a pathologist to identify cancer, fibrosis, and normal adjacent areas and imaged to create a database of 684 SHG images. Collagen signatures in each region were characterized using a multi-scale wavelet-based anisotropy technique developed in the CompuMAINE Lab. SHG images were wavelet transformed at 50 different size scales to calculate an anisotropy factor at each scale which identified statistically significant differences of collagen fiber organization between normal and cancerous regions. This method also identified unique characteristics of short and long-range collagen organization to create a potential classifier with predictive capabilities. This novel combination of SHG microscopy and wavelet analysis shows promise as a quantitative metric for understanding tissue structural changes, including possible laboratory and clinical applications in the future.

Non-Contact Image Analysis of Breathing Rate Using An Unmanned Aerial Vehicle (UAV)

Submission Type: Exhibit

Submission Category: Engineering and Information Sciences

Author(s):

Basel White

Maggie McCarthy

Dominic Kugell

Jacob Holbrook

Andre Khalil

Undergraduate Student Presentation

Faculty Mentor: Andre Khalil

Abstract: Due to their ability to travel through complex terrain efficiently, UAVs are often used in search-and-rescue operations to locate individuals who are lost and possibly injured. The necessary technology and ability to assess the subject’s vital signs without physically contacting them is lacking. To address these gaps, an image analysis technique has been developed to analyze videos and photographs taken in a non-contact manner and output accurate vital signs for an incapacitated individual. The image analysis software FIJI was used, and two scripts were developed: (1) to segment the chest area from the UAV output video of the respective patient and (2) to utilize convolution to detect the edge of the chest on this same video. These two approaches allowed for the difference in the chest area throughout the video to be quantified, allowing for the extraction of the patient’s breathing pattern as a function of time. Aspects of the device that are critical to the project’s success will be tested to ensure the reliability and durability of the methodology, such as comparing the efficiency of the two different strategies, (1) and (2), testing the accuracy of the body temperature reading, creating a comparison of a heart rate measurement from an oximeter versus from our scripts, and a drone calibration. Preliminary results demonstrate this methodology will successfully output viable vitals, adding a new aspect to search-and-rescue techniques.

Growing Digital Forest

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Kingsley Wiafe-Kwakye

Torsten Hahmann

Kate Beard

Graduate Student Presentation

Faculty Mentor: Kate Beard

Abstract: Forest dynamics, which include rapid environmental changes, natural and human disturbance, and various management objectives, interact in various ways to create a diverse and complex system in the forest. Despite prior work to better understand forest dynamics, there are still many open questions about how these complex interacting factors interconnect to build the forest ecosystem. This limitation has multiple negative consequences for the future of forests, including a limited ability to forecast how trees, and hence the Earth’s climate system, will respond to rapidly changing climate circumstances. Accurate forest ecosystem predictions will help make more informed decisions when faced with erratic changes in climate conditions.
This study presents the Digital Forest, an ontology-based framework for integrating and analyzing forest related data in various forms. A forestry domain ontology and a geographical statistical ontology make up the Ontology layer, used to create a representation of spatial and time varying forest-related data. The Ontology layer provides the integrative language for expressing comprehensive analytical questions. The SPARQL/GEOSPARQL layer provides a scalable query and exploration interface for the integrated data, while the Machine Learning layer provides tools to spot and summarize more complex trends and unknown associations. Presently, the ontology has enabled the development of a knowledge base with data on terrain, temperature, geology, and tree species that connects to the digital forest’s semantic interface to support query and visualization operations. The study showed that ontology could be used to integrate and to explore diverse spatio-temporal data sets that can provide answers to complex questions.

SMART Tourniquet

Submission Type: Exhibit

Submission Category: Engineering and Information Sciences

Author(s):

Abigail Varney

Kaisa Heikkinen

Benjamin Hutchins

Kaitlyn Webber

Lisa Weeks

Undergraduate Student Presentation

Faculty Mentor: Lisa Weeks

Abstract: Tourniquets are used in emergency scenarios to stop life-threatening blood loss, and in the process, they often cause irreversible tissue damage that can lead to acute compartment syndrome (ACS) or amputation. Certain companies have created surgical tourniquets that reduce tissue damage and blood loss, however, these devices are not effective in the military or search and rescue applications as they are not portable or durable. The purpose of this work is to create a safe and affordable tourniquet for emergency situations. A device was designed to address irreversible tissue damage. Its effectiveness was validated in four key ways: temperature variation testing of an endothermic chemical reaction, delivery mechanism testing of a topical numbing agent, pressure testing an automatic air pump, and sensor testing for heart rate measurements. The temperature variation testing results showed that there was a 50% decrease in temperature, which cools the device from 24°C to 13°C upon application. Testing indicated that delivering the numbing agent by punching holes in packets containing the agent results in an adequate amount of medication reaching the skin when pressure increases. The results suggest that our tourniquet is applicable in emergency scenarios while eliminating unnecessary tissue damage caused by current tourniquets.

Acknowledgements:
The authors thank Dr. James Beaupre, Dr. Robert Bowie, Dr. Lisa Weeks and Liza White for design and application help, and critical advice.

Microscopy Methods Second Harmonic Generation and Fluorescent Photo-activated Localization Microscopy and their Applications in Biological Systems

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Jan Wusik

Sam Hess

Karissa Tilbury

Graduate Student Presentation

Faculty Mentor: Karissa Tilbury

Abstract: The introduction of lasers into the sciences has enabled researchers to investigate many different types of systems, from atoms to biological systems. We discuss two novel microscopy methods, Second Harmonic Generation (SHG) and Fluorescent Photo-activated Localization Microscopy (FPALM) and their applications into biological imaging. We present an overview of the physics behind SHG, and discuss its biological applications. We show data from a specific study of mouse skin from an experiment where the integrin alpha-v beta-3 was knocked out, while comparing it to the wild type sample. We also captured the auto-fluorescent signal, and found the ratio of intensities between a coupled SHG image and the auto-fluorescent signal. We concluded that there was not a statistically significant difference between the two samples in terms of this ratio, meaning that the integrin alpha-v beta-3 does not have an effect on the structure of collagen.
We provide a similar analysis of FPALM, and we show a cell expressing hemagglutinin (HA) marked with photoactivatable green fluorescent protein (PA-GFP). FPALM was able to show clusters of HA in the cell in a high resolution, and track movement of HA in a live cell. Finally, we discuss the systems where they are most effective, along with their advantages and disadvantages over each other.

Identifying Valid Sizes for Raster Regions in Topological Relations – Presentation withdrawn from judging

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Brendan Hall

Graduate Student Presentation

Faculty Mentor: Max J. Egenhofer

Abstract: Topological relations are an essential element of spatial queries and reasoning about spatial information. The predominant model for topological relations in geographic information systems — the 9-intersection — identifies sixteen different relations between groups of pixels (called raster regions) in a raster space. Several of these relations are dependent on the raster region sizes to be realized. An example named ‘Completely Inside’ would require rasters to be sufficiently different in size for one raster to entirely encompass the other. Based on the size ratio required to realize certain relations, the hypothesis is that relations are predictable based on the two rasters vertical and horizontal extents. In order to determine the baseline for the realizability of each of the sixteen relations, a Python program was developed that calculates all possible relations for two fixed-size raster objects by determining their unique 9-intersection matrices. Visualizing the size-based existence of the relations over the sixteen relations’ conceptual neighborhood graph — an arrangement of the sixteen relations by highest similarity — patterns and trends become visible so that the hypothesis can be tested.

Low-Cost On-Board Computer for CubeSats

Submission Type: Virtual Presentation

Submission Category: Engineering and Information Sciences

Author(s):

Steele Muchemore-Allen

Joseph Patton

Travis Russell

Ali Abedi

Graduate Student Presentation

Faculty Mentor: Ali Abedi

Abstract: MESAT1 is a CubeSat that was proposed by the University of Maine in response to NASA’s CubeSat Launch Initiative, and in early 2020 was selected by NASA to be launched into orbit in 2022. An on-board computer (OBC) is a subsystem of a CubeSat that is designed for monitoring and control of other subsystems, processing associated with the payload, and data packeting associated with communication. The MESAT1 OBC was designed to monitor the specific subsystems of MESAT1, as well as to manage the science payload, four consumer grade digital cameras. The OBC will also be the primary form of communication with the on-board radio, a linear transponder module (LTM) provided by AMSAT. The OBC was designed around the use of a Raspberry Pi Compute Module 3+ system-on-chip (SOC). Two of these SOCs were used on the OBC, with one being the primary controlling unit and the other being used as a redundant backup. This allows the system to keep functioning after any possible single-upsets from ionizing radiation in space. The MESAT1 OBC can be produced for a low cost, while a commercial off-the-shelf OBC can cost tens of thousands of dollars.

A Customizable, Accessible, and Multimodal Smartphone Application to Monitor Speaking Rate in a Clinical Setting for Children with Fluency Disorders

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Aubree Nygaard

Jessica Holz

Christopher Dufour

Undergraduate Student Presentation

Faculty Mentor: Christopher Dufour

Abstract: For children with fluency disorders, such as stuttering disorders, research demonstrates significant negative life impacts in the areas of educational achievement as well as social, emotional, and mental health. Approaches to intervention and management of fluency disorders for children often involve the support and clinical treatment by a Speech Language Pathologist. One common strategy taught in therapy is for children to use a reduced rate of speech. This method, unfortunately, presents limitations. It is clinician dependent, which is more likely to be subjective, inconsistent, and delayed. It also may not encourage carryover to situations in which a clinician is not present. With this project we present an alternative method for teaching reduced speech rate through an objective, real-time feedback measure, presented within a smartphone application. Though some smartphone applications for speech rate monitoring exist, our application aims to improve on existing designs in its specific intended use for children in a therapeutic setting and its emphasis on accessibility, customization, and multimodal interaction. Our design enables a clinician to customize the application to suit a particular child’s needs through options including haptic, visual, and auditory cues, creating a more child-centered experience. The user interface changes in real-time as the child speaks, thereby providing them with a method to independently monitor and adjust their speech rate if needed.

Development of a Rapid LAMP-Based Field Method for Fecal Contamination Monitoring

Submission Type: Virtual Presentation

Submission Category: Engineering and Information Sciences

Author(s):

Kettie Cormier

Casey Schatzabel

Jean MacRae

Undergraduate Student Presentation

Faculty Mentor: Jean MacRae

Abstract: Rapid and field deployable fecal contamination tests are needed to ensure the safety of our drinking water, recreational water, and shellfish harvests. In this project, I developed equipment to enable field application of a loop mediated isothermal amplification (LAMP) test method to track fecal contamination in water. LAMP is a highly specific and sensitive gene amplification technique that has been used as a simple, rapid diagnostic tool for early detection and identification of microbial diseases (Parada et al., 2008). A portable heating device was designed to maintain the required temperature for the amplification reaction. Also, to make the results more reliable, and possibly quantitative, a light source and a phone-based detection system were designed to measure the color change produced in the reaction. This research contributes to the development of a simple and inexpensive test that doesn’t require a lab, and that anyone can use to test the safety of their water.

Boardman Computer Science Lab Web Portal – Presentation withdrawn from judging

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Aaron Schanck

Jack Brisson

Klei Bendo

Alex Landry

Samual Morse

Forrest Swift

Terry Yoo

Undergraduate Student Presentation

Faculty Mentor: Terry Yoo

Abstract: The Boardman Computer Science Lab is a long-standing resource for UMaine Computer Science students. Its main purpose is to aid students with issues and questions related to computer science and acts as a meeting place for students working in groups. However, the Boardman Lab has historically struggled in broadcasting accurate and up-to-date scheduling, and stagnation in terms of systemic improvements.
In-House Operations in conjunction with Mr. Christopher Dufour, the director of the Boardman lab, is producing the Boardman Computer Science Lab Web Portal to mitigate these shortcomings. This application will include features for students to view helpers’ availability dynamically and schedule help sessions, and collect data automatically to better understand where students need help as well as saving user feedback to gauge the effectiveness of specific helper’s methods.
The Boardman Web Portal is currently in development at approximately 75% completion, set to be completed and distributed for alpha testing in early April. We are, at present, able to effectively demonstrate remote student sign-in, scheduling visualization and session creation, and student feedback. In the coming weeks, we will finalize the data analysis features and other small miscellaneous and aesthetic updates.
Our project aims to mitigate the current shortcomings of the Boardman Lab, as well as to propel it towards continuous improvement. The In-House operations team is aiming to provide a genuine utility service for a program that has helped each of us greatly throughout our undergraduate experience so that future students can have an even greater experience.

Artificial Intelligence for Energy Efficiency (AI4EE)

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Noah Lambert

Drew Hooke

Undergraduate Student Presentation

Faculty Mentor: Drew Hooke

Abstract: The University of Maine Orono campus contains approximately 200 buildings, many of which contain climate-sensitive equipment with occupants coming and going all day, every day. Currently, a majority of buildings on campus rely on the most basic, conventional thermostats that are set to a specific temperature. These conventional thermostats continue to use energy until the desired temperature is reached, regardless of whether the building is occupied or needs constant temperature. Though some buildings require more consistent climate control, other buildings may not. In essence, conventional thermostats are all well and good until a building is left unoccupied with temperatures being maintained unnecessarily. Additionally, because conventional thermostats narrowly focus on meeting a set temperature, they can also fail to meet the strict climate control needs of sensitive equipment when other variables (such as weather) change and impact the internal temperature of the building. AI4EE seeks to solve these problems by creating a thermostat that harnesses the power of artificial intelligence to generate smart schedules for buildings. These smart schedules will adapt to constantly changing variables that affect climate, including the building’s current and upcoming occupancy status and the effectiveness of the building’s insulating properties. During this performance period, a custom prototype of a wireless temperature and humidity sensor was developed. These sensors were designed as a base station-satellite system (BSSS) consisting of one central computer. The central computer wirelessly manages and collects data from several satellite sensors before packaging the data to be sent for future processing. These developments highlight the massive amounts of potentially wasted energy from heating our buildings and may ultimately help pave the way for a more energy-efficient campus.

Wave Attenuation Over Submerged Oyster Farms – Presentation withdrawn from judging

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Liam Hanley

Kimberly Huguenard

Graduate Student Presentation

Faculty Mentor: Kimberly Huguenard

Abstract: This project is one of several projects currently active in the Coastal Engineering group at the University of Maine’s Department of Civil and Environmental Engineering. The purpose of this study is to examine the use of sunken, over-wintered oyster farms as a strategy to reduce shoreline erosion. Wave forces constantly erode shorelines across the world, and natural solutions to coastal erosion have been studied to be effectual and generally better for ocean ecosystems than “hard engineering” solutions. Aquaculture represents a large portion of Maine’s economy, and some studies have been conducted on capitalizing on the breadth of Maine aquaculture for engineering solutions. This project involves measuring wave characteristics on a longline of oyster cages owned by Maine Ocean Farms, in Freeport, Maine. Field observations will be used to produce a numerical model of the localized area, so that different environments and farm layouts can be tested for efficacy. This project will inform oyster farmers and policy makers on layout strategies for their sunken oyster farms, with shoreline retreat in mind. This study will quantify the effect of bottom lying oyster gear on reducing wave height from winter storms and inform how these effects can be increased generally in Maine, to reduce coastal erosion. This project is currently ongoing and will have results by the end of Summer 2022.

Building a Mobile Application to Facilitate a Safe Virtual Reality Experience – Presentation withdrawn from judging

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Nicholas Sherman

Ethan Trott

Sam Braga

Christian Doiron

Chris Vogel

Terry Yoo

Undergraduate Student Presentation

Faculty Mentor: Terry Yoo

Abstract: In recent years, VR technology has consistently improved and become more desirable to the average consumer. As a result, more companies are developing VR applications that include opportunities for people to interact online. The goal of this project is to work with Real Time Reality to provide this interaction in an application that is simple and accessible. The project includes the creation of a Progressive Web App (PWA) for use on iOS and Android devices. The PWA will allow users themselves to be modeled in virtual space, also called the “metaverse”. This “metaverse” is the focus of Real Time Reality. The goal is to create a safe and secure environment using a virtual avatar that is based on their real life appearance. This model will be created and managed through the PWA, by allowing users to create an account and sign in, either via a username and password or through biometric authentication. Users will be able to take a short video of themselves, and upload it to a server to be converted into their avatar. The PWA will show the status of the avatar model generation to the user. Then, once a user has a model, the PWA will coordinate the connection between the smartphone and the VR headset, transforming the real time video feed into pixel coordinates, sending them to the VR headset over this connection. This will facilitate the animation of the user’s model in the metaverse, allowing for real time movement.

Hybrid Carbon Nanotube based Catalytic Materials as a Chemical Sensor

Submission Type: Virtual Presentation

Submission Category: Engineering and Information Sciences

Author(s):

Bivek Bista

Sharmila Mukhopadhyay

Wenhu Wang

Undergraduate Student Presentation

Faculty Mentor: Sharmila Mukhopadhyay

Abstract: This research aims to investigate a new type of sensor platform for using hierarchical hybrid solids comprising of carpet-like arrays of Multi Walled Carbon Nanotubes (MWCNT), attached to a conducting porous carbon substrate, and activated with analyte-sensitive nano-catalysts. For this project, we are using palladium (Pd) nanoparticles (Vijwani et al 2012, and Wang et al, 2021) that interact with chemical species such as ammonia. It is expected that the MWCNT carpets activated with Pd nano-catalysts will provide a significant increase in the number of reaction sites resulting in more electron transfer between the sensor and analyte, hence show increased sensitivity as chemical detector.
Samples were synthesized using Plasma Enhanced Chemical Vapor Deposition (PECVD) and Thermal Chemical Vapor Deposition (TCVD) process. Material Characterization was performed using Scanning Electron Microscope (SEM), X-ray diffraction (XRD), and X-ray Photoelectron Spectroscopy (XPS). Two different techniques were used to evaluate the correlation between processing, structure, and electrical properties of these materials: measurement of resistance using Keithley multimeter, and electrochemical behavior using Cyclic Voltammetry process.
Three different materials have been compared: carbon fiber cloth (CFC substrate), MWCNT enhanced CFC (MWCNT-CFC), and Pd nanoparticles attached to MWCNT-CFC (Pd-MWCNT-CFC). Our study confirms that these materials, especially Pd-MWCNT-CFC samples, exhibit promising electrical response, and may potentially lead to flexible and customizable sensing platforms in the future.

Improving Coastal Water Quality Monitoring and Decision-Making Through Rapid, Simple Field Testing – Presentation withdrawn from judging

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Casey Schatzabel

Kettie Cormier

Jean MacRae

Graduate Student Presentation

Faculty Mentor: Jean MacRae

Abstract: There is a need for field-deployable methods that can quickly detect fecal indicator bacteria (FIB) to assess the safety of drinking water sources, recreational waters, and seafood harvests. Water quality has traditionally been monitored using bacterial culture-based methods, which is a time-consuming process (Jiang, 2018). Newer molecular diagnostic methods, such as PCR-amplification, require specially-trained lab workers and expensive equipment (Zhu, 2021). The goal of this research is to use LAMP (loop-mediated isothermal amplification)-based technology to develop a quicker, easier, and less expensive field-based detection method that does not require specialized equipment. A further goal is to minimize the waste produced in using this test method. The best testing protocol depends on the goals of the user, but all methods consist of three steps: sampling/concentration, extraction/preparation, and detection. Two concentration steps will be assessed in this work: grab sample filtration and passive sampling. Simple heat treatment will be compared to field-manageable extraction methods, such as paramagnetic bead or column-based nucleic acid purification, for sample preparation. The LAMP detection step will be optimized for the required level of sensitivity for specific users. The results will be compared in terms of cost, time, materials, compatibility, waste generation and ease of use. Appropriate workflows for concentration, extraction and detection of E. coli will be identified for users who prioritize different factors, such as location of testing, cost or sensitivity. These workflows may generate more efficient monitoring techniques for regulatory bodies, municipalities, aquaculture operations, and off-grid system management.

547. Facilitating Early Detection of Neuropathy

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Ashley Drexler

Jon Donnelly

Sammy LaRochelle

Maxwell Prybylo

Travis Tovey

Terry Yoo

Undergraduate Student Presentation

Faculty Mentor: Terry Yoo

Abstract: Small fiber neuropathy can be caused by over 30 medical conditions including diabetes, aging, and chemotherapy. Most common symptoms include: pain, numbness, tingling, and loss of motor functions. Currently there are no reliable options for early detection of small fiber neuropathy. Recently much work by Neuright, Inc has focused on early detection of neuropathy by measuring electrical activity in nerves via a microneedle array providing a readout on the nerve health. In this project we create a user interface for such a device so that the electrical activity can be viewed or stored. The user interface includes a robust database that connects to the microneedle array and securely receives electrical activity. The interface was designed to be deployed in a hospital setting and includes the ability to add multiple clinicians, staff, and patients. In addition, we also develop a machine learning algorithm to analyze the electrical activity for signs of small fiber neuropathy. Through these features, this project paves the way for easily and effectively implementing the early detection of neuropathy in a normal clinical setting.

3D Image Segmentation of Cells for Localized F/B SHG Analysis

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Betelhem Solomon Abay

Karissa Tilbury

Graduate Student Presentation

Faculty Mentor: Karissa Tilbury

Abstract: Early detection campaigns for breast cancer have radically increased the number of breast cancer diagnoses each year; however, the number of women dying from metastatic breast cancer is stagnant. Increased understanding of the tumor microenvironment and the cellular and acellular signaling cues associated with tumor promotion and metastasis. Here we develop a 3D breast tumor model of breast cancer cells (MCF7A or MDA-MB-231) co-cultured with fibroblasts cells (cAMP activated +/-) to recapitulate the interaction of breast cancer cells and neighboring normal fibroblasts for 72 hrs. Collagen fibers were imaged using Second Harmonic Generation (SHG) imaging in both the forward and backward directions. An image analysis pipeline was developed for pixel-wise cell segmentation and the emission directionality FSHG/BSHG ratio (sensitive to phase matching and the dipole organization in the fibrillar packing at size scale of λSHG) calculation. The analysis targeted to quantify the fibrillar organization locally at 10 pixel (3.52μm) dilation from the cell boundary. The gels with (+) cAMP fibroblasts with either MCF7A or MDA-MB-231 had lower FSHG/BSHG ratio from more packed & aligned smaller diameter fibers. The gels with (-)cAMP fibroblast with either MCF7A or MDA-MB-231 had a higher FSHG/BSHG ratio from larger diameter fibers that are less randomly packed. Both MCF7A and MDA_MB_231 cancer cells induced increased remodeling in collagen when the fibroblasts are tumor-associated than (+) cAMP fibroblasts. These tumor microenvironments were characterized by higher collagen deposition, increased stiffness, dense packing, and more alignment near the cell surface which favors tumor migration to secondary locations.

PriCon: A Privacy Conscious System

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Maxwell Prybylo

Vijayanta Jain

Sepideh Ghanavati

Undergraduate Student Presentation

Faculty Mentor: Sepideh Ghanavati

Abstract: Users often skip reading privacy policies to understand privacy behavior of web and mobile applications because these documents are long, complex, and ambiguous. Recent work has addressed these limitations by shortening the policy and creating conversational privacy chatbots to answer users’ privacy questions. Although these work provide good information to users, it may offer inaccurate information because privacy policies often misrepresent the privacy behavior of applications. In this project continuation, we plan to expand on our previous work and create PriCon: a chatbot that responds to users’ privacy questions about an application based on the privacy policy practices extracted from the privacy policy and privacy behavior captions created from an applications source code. In addition, we plan to release our corpus of privacy policies, language embedding, and trained models.

Effects of Salinity Changes on Freshwater and Moderately Saline Biofilters

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Eliza Costigan

Jean MacRae

Graduate Student Presentation

Faculty Mentor: Jean MacRae

Abstract: Fish typically excrete nitrogen in the form of ammonia, which is toxic to most organisms even at low concentrations. In recirculating aquaculture systems (RAS), where only 10% of the system water is exchanged per day, the ammonia-nitrogen produced by fish must be converted to the less toxic nitrate-nitrogen before recirculation. The microbes that perform this conversion, nitrifiers, are notoriously sensitive to changes in environment, especially changes in salinity. This can be an issue in RAS when transitioning anadromous fish such as Atlantic salmon from freshwater to saltwater over their lifetimes. However, acclimation to a small amount of salinity before transition to a higher salinity may help biofilters recover from these changes. A series of experiments was performed on both freshwater and moderately saline (3 ppt) biofilters to assess their respective levels of recovery after an abrupt change in salinity (3, 20, and 33 ppt). Tests were run for a two-week period in which the nitrification rates were monitored. It was found that the nitrification rate in the freshwater biofilters recovered by approximately 91% in two weeks after a shift to 3 ppt water, and showed no recovery after shifting to 20 and 33 ppt water. The nitrification rate in the moderately saline filters recovered by about 11% in two weeks after a shift to 20 ppt water, and did not recover at all in 33 ppt water. These results indicate that it is possible to prepare biofilters for a major, abrupt salinity shift using a low level of salinity.

Thermally Protected Phone Case – Presentation withdrawn from judging

Submission Type: Virtual Presentation

Submission Category: Engineering and Information Sciences

Author(s):

Aimee Whitman

Jon Ippolito

Undergraduate Student Presentation

Faculty Mentor: Jon Ippolito

Abstract: My capstone project idea is to design a phone case that defends devices against temperatures of the hot summers and cold winters to keep the hardware and batteries of mobile devices from being permanently damaged. Oftentimes, if a phone is left out in the cold for too long, users report that their charge drains to almost nothing in minimal time with minimal usage.

Likewise, when using a phone at the beach, leaving it in the car, etc., the phone will turn black and not be usable due to overheating. Therefore, using materials inspired by NASA spacesuits, household kitchenware, and thermally-protected clothing, I want to design a sleek, durable, and lightweight phone case that is thermally regulated on the inside.

Air Dehumidification Using CNC/PDMS Dense Membrane at Elevated Temperatures and Relative Humidities

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Nasim Alikhani

Ling Li

Graduate Student Presentation

Faculty Mentor: Ling Li

Abstract: Moisture-selective membrane technology shows great potential in conventional wood drying processes as it provides a solution to recycle the thermal energy stored in the exhaust air vented from the kiln. This study aimed to develop a high-performance cellulose nanocrystal (CNC) modified PDMS composite membrane material to efficiently separate water vapor from the moist warm air at a high temperature, up to 80°C. Three objectives were to 1) determine the optimal ratio of CNC as an additive to PDMS polymer; 2) fabricate a CNC/PDMS/PAN composite membrane material; 3) evaluate the permeability and selectivity of water vapor and Nitrogen gas at high temperatures (up to 80°C), and different relative humidity (RH) levels. Weight concentrations of CNC in PDMS were 0%, 1%, 2%, 3%, 4%, and 6%. The water vapor permeability of the membrane samples was measured with the Payne diffusion cell, coupled with the Dynamic Vapor Sorption (DVS) instrument, and the nitrogen gas permeability with a gas permeation cell. The selectivity was obtained from the ratio of water vapor permeability to nitrogen gas permeability. The results showed that the optimal addition of 2% CNC to the PDMS increases the water vapor permeability of the membrane by 24.8%, 30.9%, and 11.2% at 25 °C, 50 °C, and 80 °C, respectively. The selectivity was slightly increased by up to 7.7% when 1% CNC was added. The research findings will also facilitate the application of CNC in membrane separation fields.

Ad Hoc Positioning in Wearable Devices

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Jacob Mealey

Mike Scott

Undergraduate Student Presentation

Faculty Mentor: Mike Scott

Abstract: Over the last few years wearable devices have come into the mainstream. These devices provide many benefits to many users, but are limited as they must be tethered to a cell phone. Ad-hoc networks can be leveraged to estimate relative positioning of neighboring devices. Ad-hoc positioning systems are actively being researched, but much of the research is based on stationary networks. In preliminary work we have deployed positioning systems working on small embedded devices. In the proposed work, we will develop a variation of existing work in ad-hoc networks to apply it to wearable devices through simulations to determine optimal design specifications as well as the limits of such designs.

Immiscible Liquid-Coated Materials for Water and Aerosol Filtration

Submission Type: Virtual Presentation

Submission Category: Engineering and Information Sciences

Author(s):

Justin Hardcastle

ChunKi Fong

Danika Evangelista

Rushabh Shah

Shao-Hsiang Hung

Aydin Cihanoglu

Caitlin Howell

Graduate Student Presentation

Faculty Mentor: Caitlin Howell

Abstract: During the purification of water and air, the adhesion of bacteria and other foulants is an ongoing issue that leads to flux decline, and current chemical and physical cleaning methods for fouled filters can lead to filter degradation over time. In addition to being porous throughout, commonly used filtration materials such as polytetrafluoroethylene (PTFE) and polyvinylidene fluoride (PVDF) often have a textured surface that promotes the adhesion of bacteria and other contaminants. In this work, we present bioinspired liquid-coated filters as a new approach to creating filters that resist fouling. Liquid-coated filters were created by immobilizing a water immiscible liquid on the surface of commercially available synthetic filters. For water filtration tests, 0.45 µm pore diameter PTFE and PVDF filters were coated with omniphobic perfluoropolyether liquids. We measured the continuity of the surface liquid layer by testing how easily a water droplet could begin to move the surface, as well as the anti-adhesion properties of the surface liquid layer through quantifying the speed of a droplet’s movement at a static angle of inclination. The results indicate that infused PTFE membranes form a superhydrophobic surface with a sliding angle of approximately 5ᵒ. Infused PTFE membranes sustained a more consistent functional liquid layer, with an approximately 75% lower sliding angle and 70% faster droplet movement at a fixed angle, than the infused PVDF. We measured the ability of the infused membranes to resist biofilm formation by incubating in growth media with Staphylococcus epidermidis for 24 hours. Infused PVDF membranes reduced biofilm formation by approximately 25% compared to bare controls while infused PTFE membranes reduced biofilm formation by approximately 98%. Pure water permeability (PWP) experiments conducted at an applied pressure of 1.5 bar indicated that liquid-coated PVDF membranes had a statistically equivalent PWP of 2827 ± 323 L/m2-h-bar, for over 10 cycles of use, showing that the immobilized liquid is present and stable within the pores. For aerosol filtration, we tested the efficiency in capturing and releasing aerosolized Escherichia coli using liquid-coated commercial PTFE and HEPA filters. We determined that bacteria trapped on the liquid-coated filters could be removed with increased efficiency compared to bare controls. The use of liquid-coated materials in water and air purification applications opens new doors for the creation of a biointerface that resists adhesion in dynamic environments.

555. Equiptrac: Solving Equipment Management Time Usage and Cost Reporting for the ASCC – Presentation withdrawn from judging

Submission Type: Exhibit

Submission Category: Engineering and Information Sciences

Author(s):

Tyler Delargy

Michael Rumohr

Michael Taylor

Brian Couture

Spencer Campbell

Terry Yoo

Undergraduate Student Presentation

Faculty Mentor: Terry Yoo

Abstract: The purpose of this product is to provide equipment usage tracking and cost accounting capabilities to the ASCC. Users of this software will be both center employees, as well as administrators responsible for equipment, lab safety, and equipment cost reporting. Employees of the ASCC constantly use expensive lab equipment for various types of testing and manufacturing. Employees may be working on one or more projects, for many different high- profile clients from national labs to public and private companies and government contracts. Given that equipment has an expected lifespan and associated cost, in order to properly charge ASCC clients for work done using equipment, an equipment usage tracking system is necessary. Employees will use the created software to log date, time, and duration of use of a specific piece of equipment tied back to their name and project number they are working on. Administrators will use the software to update equipment database information and generate reports of equipment usage over various time scales and filters. In response to this problem, we have created a mobile and web application that can be accessed and used to sign in using UMaine single sign on, then check their current equipment in use, make a new entry, generate usage reports, and then access administrator functions if allowed. We also implemented an additional function beyond our requirements, a machine learning model for text recognition to automate the process of making a new equipment usage entry.

Starting Fresh! Gathering New Information on Freshwater Flows Pouring into Maine’s Estuaries

Submission Type: Virtual Presentation

Submission Category: Engineering and Information Sciences

Author(s):

Hanna Cronin

Sean Smith

Neil Fisher

Undergraduate Student Presentation

Faculty Mentor: Sean Smith

Abstract: Coastal water quality and aquatic habitat conditions are largely driven by fresh and saltwater inputs to estuaries. Information on the freshwater flows is essential for detection and prediction of problems related to coastal ecosystems, seafood industries, and public health. The monitoring and prediction of stream flows can be challenging due to complicated measurement logistics and modeling uncertainties. Here we address these challenges in the Frenchman Bay region using an ensemble of measurement approaches, followed by comparison of stream discharge time series to watershed rainfall runoff simulations. Monitoring flow in Crippen’s Brook in Trenton, Maine allowed for the development of a discharge time series that is compared to the three different forecast ranges of the National Water Model (NWM) using the Nash-Sutcliffe efficiency evaluation. The measurement and post-processing of the discharge time series in the study watershed, provides a new fresh start for quantification of runoff inputs into the Bay. The flow time series provides important information for new modeling of land-sea connections and estuary hydraulic conditions, as well as a basis to expand uses of the NWM for land-sea connections affecting estuaries in the future. Results from analyses of individual storm events and extended time series indicate that the NWM is currently a relatively poor predictor of runoff (freshwater flows) in the Frenchman Bay watershed with the longest-range prediction producing the most accurate outcome. Under-prediction of flows by the NWM when compared to our measurements can be attributed to uncertainties in the meteorological and runoff components of the NWM simulations.

DeepReal: Exploring Methodology for 3D Streaming of Authenticated Humans – Presentation withdrawn from judging

Submission Type: Exhibit

Submission Category: Engineering and Information Sciences

Author(s):

Tyler Delargy

Terry Yoo

Undergraduate Student Presentation

Faculty Mentor: Terry Yoo

Abstract: Real Time Reality is a student founded and owned startup looking to revolutionize digital interaction online. With the goal of making socialization on the internet grounded in reality, authentic, private, and secure a new technology is needed to stream authenticated real people into virtual environments for web3 and beyond. To do this, a pipeline was investigated for creating realistic true to life avatars for any person using mobile hardware. These replicas can be used as both a representation of virtual identity, and a biometric key to keep information private and secure and will be stored and managed via decentralized ledgers to insure no central control over this shared and authenticated reality. Novel combinations of methods were used to predict human body shape from several measurements and images, however, most of the investigation was into the human face and head where humans identify individuals. Preliminary results show efficacy of traditional photogrammetry in this task; however, this method produces inconsistent topology unable to be rigged for animation or stored easily as a modification of a base mesh. Therefore, further methods to use shape from shading, facial key point detection networks, human segmentation priors, and training of deep learning GAN networks are being investigated create a process that can be performed by anyone with a mobile phone to produce consistent and realistic results.

The Utilization of LIDAR Technology for Realistic VR Simulation – Presentation withdrawn from judging

Submission Type: Virtual Presentation

Submission Category: Engineering and Information Sciences

Author(s):

Ethan Woods

Michael Scott

Undergraduate Student Presentation

Faculty Mentor: Michael Scott

Abstract: While VR is rapidly becoming a more mainstream medium for sharing experiences in a way never before seen, LIDAR technology has existed for as long as the practice of taking terrain data left the ground. With the ability of VR to provide lifelike interaction and clarity to anything imaginable, simulation has always been one of its primary utilizations, especially in the education and industry sectors. When it comes to a project in which the goal is to immerse somebody in a realistic and accurate environment mirroring that of the real world, the combination of LIDAR and VR is an attractive mixture. With these two technologies, we are working on an experience in which you are transported to a remote lake in Alaska and are get to experience and engage in the environment around you as if it were the real thing. Specifically, we are simulating this environment thousands of years ago, when animals such as mammoths still roamed its tundra. Considering the primary goal of this project is to make this experience available to kids to learn more about the process of Paleontology and the work of a Paleontologist, realism is key. Therefore, with the use of LIDAR and other technologies, we will be able to provide the most authentic experience possible, instilling in our users that what they are seeing is, in fact, real-life, and a real profession that they are engaging in.

Relay Assisted Wireless Energy Tansfer for Efficient Spectrum Sharing in Harsh Environments

Submission Type: Poster

Submission Category: Engineering and Information Sciences

Author(s):

Sonia Naderi

Ali Abedi

Graduate Student Presentation

Faculty Mentor: Ali Abedi

Abstract: Since the number of wireless sensing devices are increasing rapidly, the concept of energy efficiency is of high importance and will lead to high spectrum efficiency. One of the main hurdles in efficient spectrum sharing in wireless sensor networks is the problem of power management at the node level to promote longevity without polluting the spectrum, while promoting collaboration. Wireless energy transfer for powering sensor nodes in wireless sensor networks has drawn significant research attention recently especially in hard to reach areas like forests and space. Modeling the charging and power utilization processes can help with smart transmission decisions, which can eliminate unnecessary transmissions. We are considering a cooperative relay assisted system model consisting of a transmitter powered through an energy source by both direct and relay links. An energy efficient scheduling method based on artificial intelligence is proposed for the system model to determine whether to transmit data or stay silent based on the stored energy level and channel state. All theoretical results are validated by numerical simulations and verify the effectiveness of energy relaying and proposed energy efficient scheduling methods in reducing the outage probability of the system. In this work, it is proposed that the transmitter is powered using an energy source. For future work, we will consider several batteries as an energy source powered by Photovoltaic solar panel. It is clear that by increasing the number of batteries and size of the solar panel, the outage probability of the system will be decreased but it leads to increase in the cost of the system. As a result, we will solve the problem of outage probability of the system based on the number of batteries and size of the solar panel and find the optimal value of these characteristics for a defined tolerable outage probability of the system.