Md Abdul Baset Sarker, a PhD Candidate in Electrical and Computer Engineering at Clarkson University, was selected among the Top 3 presenters (awarded $150) at the IEEE Sensors in Spotlight 2025, a global forum hosted by the IEEE Sensors Council to spotlight outstanding student-led innovations in sensor technology.
Sarker presented his work titled “AI on Edge Device: A Vision-Aided Prosthetic Hand for Children,” which proposes an affordable, intelligent prosthetic solution for children aged 10–12 years with upper-limb reduction. Unlike EEG- or EMG-based prosthetic systems that often require extensive personalized training, this novel approach uses computer vision and deep learning to simplify user interaction while enhancing functionality.
Key features of the system include:
- Vision-based object detection using a YOLOv7 model, achieving 96% mAP,
- Force and grasp classification with around 100% grasp recognition accuracy and low force prediction error (MAE = 0.018),
- Gesture detection through a CNN-based model with 99–100% accuracy,
- A lightweight (<500g), sensor-rich hand powered by an FPGA (Xilinx KR260) for real-time, low-power performance.
The prosthetic integrates a wrist-mounted camera, five force sensors, and eleven IMUs to provide adaptive, intelligent control of three degrees of freedom through servo motors. The project was funded by Clarkson University’s Center for Rehabilitation Engineering, Science & Technology (CREST) and supervised by Prof Masudul Imtiaz and Kevin Fite. It represents a major step toward accessible assistive technology for pediatric amputees.
Sarker extends his gratitude to co-authors as well as the AI Vision, Health, Biometrics, and Applied Computing (AVHBAC) lab and the Center for Rehabilitation Engineering, Science & Technology (CREST) for their invaluable support and funding.