UrbanFlow: A Hybrid Vision-LiDAR Retrofit Architecture for Autonomous Wheelchair Navigation in Complex Indoor Environments
Kofi Nyarko, David Nyarko, Derrick Cook, Charles Dankwa, Michael West, Mansoureh Jeihani
2025 13th International Conference on Control, Mechatronics and Automation (ICCMA), Paris, France
This work presents a hybrid autonomous navigation system designed for large, dynamic environments such as airports.
The system combines classical computer vision techniques for color-based track detection with deep learning-based
instance segmentation using a YOLOv8 single-shot detector, enabling end-to-end autonomous wheelchair navigation
across entire airports without user intervention.
ELT: Easy Label Trainer
David Nyarko, Kelechi Nwachukwu, Anjolie Anthony, Kofi Nyarko
Springer Nature, Lecture Notes in Networks and Systems
This paper presents ELT-Easy Label Trainer (ELT), a tool designed to automate and streamline
the process of labeling and training object detection models. It integrates advanced models like
YOLOv8, YOLOv9, and YOLO-World for robust object detection and segmentation using descriptive text prompts.
Finding Indicators of Nuclear Activity in Open-Source Images using AI/ML
R Alvarez, A Gazda, A Sarkar, David Nyarko, JJA Hastings, P Sivasankar
AGU Fall Meeting Abstracts 2023, IN33C-0735
An interorganizational approach to increase nuclear safeguards through AI/ML analysis of
open-source satellite imagery for detecting indicators of nuclear activity.
Leveraging Synthetic Image Data for Real-World Applications
R Alvarez, JJA Hastings, A Gazda, A Sarkar, P Sivasankar, David Nyarko
AGU Fall Meeting Abstracts 2023, GC21F-0976
Enhancing nuclear safeguards and capabilities through AI/ML by leveraging synthetic image data
for real-world applications in security and monitoring systems.
FDL 2023 Research Contributions
David Nyarko and SETI Institute FDL Team
SETI Institute & U.S. Department of Energy
Research contributions to the Frontier Development Lab (FDL) 2023 program, focusing on applied
artificial intelligence and machine learning for space science applications. Recognized for
outstanding performance and achievements.
edge-tpu-silva Python Library
David Nyarko (Creator and Maintainer)
Open-Source Python Library for Edge Computing
A Python open-source library that helps run large models on Edge Computers, enabling efficient
execution of computer vision tasks on resource-constrained devices. Pioneered efficient AI deployment
for edge computing applications.
