Research Symposium 2020 - Engineering | University of Detroit Mercy

E1.) A Compositional Approach to Abstraction for High-level Control of Multi-Agent Systems

Authors: Juliana Vilela (UDM), Richard Hill (UDM)

Abstract: A new approach is developed that allows abstracted graph models to be generated in a more efficient component-wise manner. This result allows the high-level cooperative control of multi-agent systems using the resulting abstracted models.

E2.) A Robust Robot Platoon for Disaster Management

Authors: Melvin P Manuel (UDM), Mariam Faied (UDM), Mohan Krishnan (UDM), Mark Paulik (UDM)

Abstract: This paper proposes a novel vision-based robot platooning algorithm that is capable of manuevering teams of search and rescue robots (SAR's) around in a dynamic disaster site under the assumption of no available communication network

E3.) Application of Fixed Colored-Target Tracking and Obstacle Avoiding Robot: Based on HS12 Microprocessor

Authors: Giovani Olivo (UDM), Joseph Afolabi (UDM) Susan Magana (UDM)

Abstract: An autonomous robot was developed as the final project in an embedded systems course. The goal was to modify a hexapod chassis by adding program code, microcontroller, camera, and obstacle sensors to search for a colored object while avoiding obstacles.

E4.) Automated Path Tracking Robot Based on Pixy2 Camera

Authors: Jiajun Chen, Changyu Jiang, Kun Chen, Mohammad Fanaei (UDM)

Abstract: The objective is to design the mobile robot to find the entrance to the course and follow the path in a specified road while detecting and avoiding obstacles.

E5.) Automatic Risk Screening System using Machine Learning and Computer Vision Approaches

Authors: Jordan Yono (UDM), Ravalika Gundu (UDM), Mina Maleki (UDM)

Abstract: The project is a COVID-19 screening system that uses computer vision and machine learning to identify risk of people entering buildings.

E6.) Design and Development of a Large Concrete 3D Printer

Authors: Bryce Sturgis, Peter Swanney, Luhan Zhang, Nassif Rayess, James Lynch (UDM)

Abstract: A Concrete 3D Printer is being designed and built that will enable students at Detroit Mercy to develop their 21st Century skills in additive manufacturing of buildings and other concrete structures.

E7.) Design of a Custom Force Fixture to Fit Miniature Load Cells for Individual Finger Force Measurement

Authors: Nichole Kutchey, Lauren Ross, Megan Conrad (UDM), Nassif Rayess (UDM)

Abstract: This project aims to develop a grip fixture measuring orthogonal forces exerted by individual finger segments. The device allows measurement of individual finger force contribution to hand strength. Ultimately, it will assess hand function in patients with impaired hand strength

E8.) Fuzzing CAN Messages using MATLAB Analyzer

Authors: Mohamad Ali Mokhadder, Samar Bayan, Utayba Mohamad (UDM)

Abstract: CAN bus messages are sniffed through Vector Tools and analyzed to determine the action instants and the primary message IDs causing the detected actions. Reverse engineering is also applied to ensure the validity of recognized IDs

E9.) Grocery Rescue: Food Recovery Application

Authors: Alessandra Pontoni, Nicholas Ibegbu, Mikeala Mansour, Nassif Rayess (UDM)

Abstract: Grocery Rescue is a food recovery mobile application intended to streamline the food donation process to help reduce hunger and promote food recovery.

E10.) Intersection Mobility for Polaris Car

Authors: Amna Mazen, Mariam Faied, Mohan Krishnan, Utayba Mohammed, Mark Paulik (UDM)

Abstract: Driving safety is expected to be improved by autonomous vehicles (AVs) compared with driving vehicles by humans. This work proposes two algorithms to guide Polaris car to turn right and left autonomously on road intersection.

E11.) Kissing Bugs Identification and Classification using Convolutional Neural Network with Transfer Learning

Authors: Bassam Abdelghani, Amna Mazen, Shadi Banitaan, Mina Maleki (UDM)

Abstract: Chagas disease classification is so important, different species don't have the same danger level, here we classify 39 Brazlizin species, 12 Mexican species, using CNN and transfer learning, then create a web app to show our work

12.) Machine Learning Techniques to Evaluate Whether Twitter Accounts Are Human or Robot

Authors: Shadi Bani Taan, Jordan Yono, Antonio Segura, Yongqiang Sun

Abstract: We use machine learning techniques to evaluate whether or not Twitter accounts are humans or robots depending on their recent activity.

E13.) Multi-Stage CNN-Based Monocular 3D Localization and Pose Estimation

Authors: Ali Babolhavaeji , Mohammad Fanaei (UDM)

Abstract: This project proposes a model for 3D object detection from 2D images taken by a monocular camera. It combines the estimated bird’s-eye view map and the deep representation of the object’s features. datasets: A syntactic dataset and the KIITI dataset.

E14.) Pin-Fin Modelling of Open-Cell Metal Foam

Authors: Omer Saad, Nihad Dukhan (UDM)

Abstract: A numerical study is performed to develop and design a pin-fin model of open-cell foam to represent the actual metal foam's internal structure, using the Ansys (Fluent) CFD method to obtain the pressure drop, permeability, and form drag coefficient.

E15). Study of Evolution By Automating Hardy Weinberg Equilibrium With Machine Learning Techniques In TensorFlow And Keras

Authors: Karthika Balan (UDM), Michael Santora (UDM), Mariam Faied (UDM), Victor Carmona (UDM)

Abstract: This research project integrates Ecology and Deep Learning by automating a principle that predicts evolution termed as the Hardy Weinberg principle using ADAM Optimization technique. The neural network

E16.) UDM Cyber-security Testbed

Authors: Samar Bayan, Mohammad Ali Mokhadder, Utayba Mohamad (UDM)

Abstract: UDM testbed is initiated within federated testbed cloud FCTaaS in collaboration with University of Arizona to test cybersecurity. Multiple attacks are tested on UDM vehicular BreadBoard within a network simulation using Vector Tools.