Michael Evans

I'm a recent Computer Science graduate & research assistant at the ODU Vision Lab working under Dr. Khan Iftekharuddin. Incoming MSE-AI online student at the University of Pennsylvania.

Previously at Old Dominion University I worked on a project involving large language models on the task of scientific claim verification, where I was advised by Dr. Jian Wu as part of the WS-DL research group.

At Lawrence Tech and Michigan State University I worked on a joint research project on developing self-drive algorithms for electric vehicles and designed a V2X software architecture as part of the CS & AI Robotics Lab. Co-advised by Dr. CJ Chung and Dr. Josh Siegel.

Research

My research interests involve task and motion planning, natural language processing, computer vision, and machine learning.

A Roadside Unit for Infrastructure Assisted Intersection Control of Autonomous Vehicles

M. Evans, M. Machado, R. Johnson, L. Escamilla, A. Vadella, B. Froemming-Aldanondo, T. Rastoskueva, M. Jostes, D. Butani, R. Kaddis, C. Chung, and J. Siegel
IEEE EIT, 2025,

[Paper]

code / video / paper / slides / pdf /

Designed a V2X wireless communication architecture with a roadside unit capable of dynamically adjusting vehicle speed in response to traffic states.

Evaluating Low-Resource Lane Following Algorithms for Compute-Constrained Automated Vehicles

B. Froemming-Aldanondo, T. Rastoskueva, M. Evans, M. Machado, A. Vadella, L. Escamilla, R. Johnson, M. Jostes, D. Butani, R. Kaddis, C. Chung, and J. Siegel
IEEE AIRC, 2025,

[Paper]

code / video / paper / slides / pdf /

Developed 5 lane-following algorithms using ROS, Scikit-learn, and OpenCV, tested on 2 GEM electric vehicles.

MSVEC: A Multidomain Testing Dataset for Scientific Claim Verification

Michael Evans, Dominik Soós, Ethan Landers, Jian Wu
ACM MobiHoc, 2023,

[Paper]

code / paper / slides / pdf /

Built a dataset of 200 scientific news claims and accompanying research papers and tested it against the gpt-3.5-turbo model on the task of scientific claim verification.

Other Projects

These include coursework, side projects and unpublished research work.

Poisonous vs Edible Mushroom Classification

Michael Evans and Grant Fitch
CS 422 Introduction to Machine Learning Term Project, 2024,

[Project]

code / slides / pdf /

Trained a gradient boosting classifier model on 47,051 samples to perform binary classification on numerical mushroom data. Achieved a precision, recall, and F1 score of 0.99 through hyperparameter tuning and 5-fold cross validation.

MARCBot IV

Michael Evans
ODU Vision Lab, 2024,

[Project]

video / slides /

Applied model fine-tuning to the VGG16 convolutional neural network (CNN) for target identification with MARCBot IV: a surveillance robot capable of identifying, tracking, and following individuals with PyTorch. Improved the motion planning algorithm in MATLAB for target following through tight, cluttered environments, enabling safe indoor operation previously limited to outdoor use.

Design and source code from Leonid Keselman's Jekyll fork of Jon Barron's website