Author(s): Raymond Magambo; James Chambo; Martine Kusekwa; Richard Simon; Happiness Kingi
Advisor: Curtis Abel
Category: Graduate
Abstract/Description: A mobile device application and web based platform which allows bus passengers to book and purchase their bus tickets directly using their mobile money accounts. Tiketi Rafiki will allow people to use their mobile phones, computers and tablets to book and purchase bus tickets anywhere at their convenience. In order to serve the offline users, Tiketi Rafiki will have agents to offer the service in different parts of the country. Tiketi Rafiki will revolutionize and improve the efficiency of the way people book bus tickets in Tanzania.
Abstract/Description: These projects are a response for the COVID-19 pandemic and address two areas: patient ventilation and medical imaging. The author designed and built a ventilator based on a BVM/Ambu bag and a robotic mechanism which is safe and ready for affordable mass production. The device provides basic ventilation modes and can be a temporary alternative to high-end equipment in a global ventilator shortage. Its design was discussed with doctors and mechanical ventilation specialists. Author’s work is fully voluntary and will be available OpenSource for any party interested in producing the device. Second project: (in progress)- Recent worldwide clinical practice has shown the importance of ultrasound for COVID-19 screening. Its is affordable, fast and shows lung structure changes at early disease stages. Unfortunately, it does require on-site ultrasound specialists and a direct contact with patients. The author is working on a robotic ultrasound system for teleoperated lung ultrasound scanning which will be tested and manufactured in Poland and then used in a hospital in Nigeria. The device allows less-qualified personnel to connect with a specialist in a remote location and also reduce the overall exposure to the virus. The specialist controls a robot that has an ultrasound probe which slides over the patient surface. This project is sponsored by the WPI Medical FUSION Lab.
Abstract/Description: Each year over 1.3 billion tons of food is thrown away, with over 90% ending up in landfills. Through the process of hydrothermal liquefaction (HTL) this waste can be harnessed and converted into a usable energy product. The addition of a bi-functional catalyst has the potential to improve the amount of energy recovered from food, thereby increasing commercial viability. This technology works to produce renewable energy from waste, helping to obtain a sustainable future.
Abstract/Description: The occurrence of hatespeech has been increasing. One of the reasons is that social media is gradually replacing traditional media as the main source of information for many people. It is very easy to reach a large audience quickly via social media, causing an increase of the temptation for inappropriate behaviors such as hatespeech, and potential damage to social systems. In particular, hatespeech interferes with a civil discourse, and turns good people away. Furthermore, hatespeech in the virtual world can lead to physical violence against certain groups in the real world, thus should not be ignored on the ground of freedom of speech. To detect hatespeech, researchers developed human-crafted feature based classifiers, and proposed deep neural network architectures. However, they might not explore all possible important features for hatespeech detection, ignored pretrained language model understanding, or proposed uni-directional language models by reading from left to right or right to left. Recently, the BERT (Bidirectional Encoder Representations from Transformers) model has achieved tremendous success in Natural Language Processing. However, even the pretrained BERT-base is heavy and expensive in the inferencing phrase. Hence, in this work, we propose to pretrain a Hateful Language Model (HLM) that (i) better understands the hatespeech related language model by pretraining from the scratch a HLM from a hateful corpus, and (ii) is light-weight by utilizing the hypercomplex space. As a result, our HLM contains only 18M of parameters compared to 110M of parameters in BERT-base model with a significant better performance compared to the fine-tuning BERT-base on hatespeech datasets.
Abstract/Description: In our daily lives, we use different kinds of paper and tissue products. Considering that 10 % of annual energy in manufacturing industries is consumed by paper drying industry, it is extremely critical to improve the efficiency of the process. Therefore, the current project is devoted to fundamentally understand the water removal from paper cellulose fibers in order to decrease the energy consumption.
Advisors: Jianyu Liang; Diran Apelian; Brajendra Mishra; Richard Sisson
Category: Graduate
Abstract/Description: In this project, we created an effective sorting, chemical composition, and composition adjustment process for metal scraps. The team has established additive manufacturing (3D printing) technology-enabled investment casting using waste steel and turned the scraps into useful high-quality parts.
Abstract/Description: Detecting depression, a prevalent and costly issue, is vital for society. As such, I have trained machine learning models to detect depression from crowd-sourced text messages with an F1 score of 0.81. Given the private nature of texts, I’m working on generating anonymous text messages to make this a sustainable solution.
Authors: Makayla D’Amore; Johann Bradley; Robert Connor; Rory Sullivan
Advisors: Holly Ault; James Hanlan
Category: Undergraduate
Abstract/Description: The Amager Resource Center (ARC) is an inter-municipal waste management company operating in Copenhagen, Denmark and its surrounding municipalities. Recently, Copenhagen has decided that the waste collection process in the city will be taken over by ARC starting in 2022 using e-trucks. These electric waste collection vehicles benefit the environment and workers as they are quieter, more efficient, and prevent the employees from breathing in harmful gas. Our group worked to design E-truck fleet recommendations for ARC specifically focusing on reducing capital investment through minimizing travel duration and distance traveled per E-truck.
Abstract/Description: This experiment analyzed the effect of subcritical water hydrolysis on natural poultry feathers for the isolation of high-value amino acids. Subcritical water hydrolysis is the process by which water in the range of 100 – 374 C converts and breaks down biomass into micromolecules. Often this process can result in the production of high-value products from poultry feathers. Poultry feathers are approximately 90 percent protein, which suggests that poultry feathers can be used to produce valuable amino acids. A non-toxic, economically viable, and environmentally favorable way to break down these poultry feathers is by subcritical water hydrolysis. In particular, this experiment focused on the effect of temperature on the hydrolysate. The hydrolysis machine was operated at three different temperatures (210, 230, and 250 C) at a constant flow rate of 10 mL / min. Differences between the hydrolysates were defined by performing Total Nitrogen, Chemical Oxygen Demand, and Nelson-Somogyi method testing. Initial hydrolysate analysis testing suggests that a low operating temperature and an operating time below 10 minutes may assist in getting optimum protein levels. No amino acid testing was conducted. All amino acid related results will be available at a later date.
Abstract/Description: To alleviate modern society’s reliance on plastic, our team explored compostable alternatives to petroleum-based polymers. This study examines the processability and properties of organic fiber-reinforced composites, constructed using a combination of PLA or a potato starch-based matrix, and natural fibers such as flax, cotton, or jute. To characterize the resulting material, our team studied the matrix-fiber adhesion, chemical composition, mechanical strength, and biodegradability of the samples. This investigation allowed our team to infer potential applications for our product, as well as issue guidance for future exploration.