Connor Blais, Graduate Student
Department of Electrical and Computer Engineering Science
Clarkson University
Wednesday, April 23, 2025
4:00-5:30 PM
CAMP 120B
ABSTRACT
This talk presents the design and implementation of an AI vision-controlled orthotic hand exoskeleton to enhance assistive capabilities for individuals with hand impairments. The system integrates a lightweight design with a deep learning model for real-time object detection. It uses the Google Coral Dev Board Micro, which houses an Edge TPU, to achieve an inference time of 52ms. Furthermore, the object detection sequence runs at 10 frames per second. The exoskeleton employs a commercially available pneumatic glove, which is controlled by a custom motor controller embedded in a custom PCB. The efficient design of the custom PCB and the codebase allows for a runtime exceeding 8 hours with a 14.4Wh LiPo battery. FreeRTOS manages multitasking for object detection, motor control, and power monitoring. Despite successes, the system faces limitations in model robustness under varying lighting conditions and object recognition accuracy. These are areas for future improvement. This work demonstrates the feasibility of combining AI vision with wearable robotics to create intuitive, user-friendly assistive devices, offering a foundation for further advancements in intelligent rehabilitation technology.
BIO
Connor Blais earned his Bachelor’s in Electrical Engineering from Clarkson University, graduating in May 2024. He is pursuing a Master’s in Electrical and Computer Engineering at Clarkson University, focusing his thesis on AI-Vision Controlled Orthotic Hand Exoskeleton. His past projects include the development of advanced Optical Particle Sensors and wearable, low-power air quality sensors at Telos Air, a startup in Potsdam.
Master’s Dissertation Committee
- Dr. Masudul Imtiaz (Research Advisor)
- Dr. Abul Khondker
- Dr. Ajay Sonar