Ultrasound Based Hand Motion and Force Estimation Powered by Deep Learning for Human-Machine Interfacing
Abstract: Biosignal-based human-machine interaction is an exciting area of research that focuses on harnessing signals from the human body to estimate hand movements. This research is crucial for designing effective human-machine interaction systems, which can control robots, augmented/virtual/mixed reality interfaces, and digital media. Ultrasound data from the forearm provides valuable insights by visualizing a cross-section of the forearm, revealing the underlying causes of hand movements and interaction. In this presentation, first, I will discuss my research on using forearm ultrasound data for hand gesture estimation. I will explore how deep learning can enhance this approach to estimate finer finger movements and forces. Next, I will present my work on making the forearm ultrasound data acquisitions more wearable, in addition to my work on developing systems and pipelines that integrate biosignal-based estimations with robotic systems and AR/VR interfaces. Lastly, I will outline future directions, including efforts towards miniaturization and system integration.
Advisor: Haichong Zhang
Committee: Robert D. Howe, Toshiaki Koike-Akino, Ziming Zhang, and Haichong Zhang
Location: Zoom - https://wpi.zoom.us/j/95671130041