Date of Award

Summer 8-16-2024

Level of Access Assigned by Author

Open-Access Thesis

Degree Name

Master of Science (MS)

Department

Computer Engineering

Advisor

Vikas Dhiman

Second Committee Member

Vijay Devabhaktuni

Third Committee Member

Prabuddha Chakraborty

Abstract

Omobot, a cost-effective autonomous mobile robot designed for indoor environments, focuses on fall detection and response for elderly care. The thesis presents the design, development, and evaluation of Omobot’s potential for aiding elderly people by combining mechanical and electronics design, software development, and improving fall detection accuracy. The robot’s design includes 3D printed structure and Mecanum wheels that allow it to move easily in indoor environment. It is powered by NVIDIA’s Jetson Nano, which helps it to process data from sensors in real-time for navigation and fall detection. The software framework built on Robotic Operating System (ROS) helps in sensor integration, real-time data processing, and provide an open-source platform. System identification modelling is used to simplify control system complexity. The system includes a fall detection system that uses images captured from the robot’s camera perspective and improves detection accuracy through homography transformation that simulates human sight levels. It uses the YOLOv8-Pose, which is a single-stage detector model, to detect the presence of humans, track movements, and identify falls. Once a fall is detected, the system sends emails to designated responders. Omobot surpasses traditional fall detection systems in accuracy, usability, and reliability, demonstrating significant potential for real-world application. The robot’s ability of autonomously navigation contributes to safer living environments for the elderly. The open-source design of Omobot ensures it can be continually updated and improved by the research community for future assistive technologies. The low-cost build can aid in removing the economic barriers that typically limit the deployment of advanced assistive technologies without compromising reliability.

Download

Files over 10MB may be slow to open. For best results, right-click and select "save as..."

Share