Augmented Reality-assisted Human Robot Interaction

January-June 2019

Final Year Project for the MEng. Electronic & Information Engineering degree supervised by Professor Yiannis Demiris.

Summary:

• Developed an augmented reality application for the Microsoft Hololens.
• AR application designed to be used by powered wheelchair users (PWUs) as an aid for navigation.
• Detects people walking in surroundings.
• Calculates a world position of the detected people to build a map of surroundings.
• Renders holographic visualizations to alert user to presence of people.
• Using the mapping, we implement a simple collision avoidance system for powered wheelchair.

Contributions:

• ROS integrated Darknet Framework, YOLO object detector trained on CrowdHuman dataset (GitHub).
• Yet Another CrowdHuman Tracker (YACHT) ROS package, tracks and determines direction of people (GitHub).
• Unity Hololens Camera video stream over ROS topics (GitHub).

Media:

• Project Report.
• Presentation.

Home | Instagram | CV | GitHub | LinkedIn