Skip to main content

Innovations for touch in VR

Sponsored by

This event is over.

Event Details:

“Deep learning classification of touch gestures using distributed normal and shear force”

Speaker: Hojung Choi, Ph.D. Candidate in Mechanical Engineering, Stanford University

Abstract: When humans socially interact with another agent (e.g., human, pet, or robot) through touch, they do so by applying varying amounts of force with different directions, locations, contact areas, and durations. While previous work on touch gesture recognition has focused on the spatio-temporal distribution of normal forces, we hypothesize that the addition of shear forces will permit more reliable classification. We present a soft, flexible skin with an array of tri-axial tactile sensors for the arm of a person or robot. We use it to collect data on 13 touch gesture classes through user studies and train a Convolutional Neural Network (CNN) to learn spatio-temporal features from the recorded data. The network achieved a recognition accuracy of 74% with normal and shear data, compared to 66% using only normal force data. Adding distributed shear data improved classification accuracy for 11 out of 13 touch gesture classes.

“AI-enhanced electronic skin that rapidly reads hand tasks with limited data”

Speaker: Kyun Kyu (Richard) Kim, Postdoctoral Scholar, Chemical Engineering, Stanford University

Abstract: Technologies for human-machine interface plays a key role in human augmentation, prosthetics, robot learning, and virtual reality. Specifically, devices for tracking our hand enables a variety of interactive and virtual tasks such as object recognition, manipulation and even communication. However, there remains a big gap in capabilities compared to human in terms of precision, fast learning, and low-power consumption.

In this talk, I explain the newly developed fast-learnable electronic skin device which enables user-independent, data-efficient recognition of different hand tasks. This work is the first practical approach that is both lean enough in form and adaptable enough to work for essentially any user with limited data.  It consists of direct printable electrical nanomesh that is coupled with an unsupervised meta-learning framework. The developed system rapidly adapts to various users and tasks, including command recognition, keyboard typing, and object recognition in virtual space.

4 people are interested in this event


Stream Information: