Author: Greg Gutmann
Affiliation: Tokyo Institute of Technology
I demonstrated my network-based VR technology at the Tokyo Nano Tech conference on January 29th to 31st with the other members of the Artificial Molecular Muscle Project. The network VR mentioned is a continuation of my work from my last paper titled Predictive Simulation: Using Regression and Artificial Neural Networks to Negate Latency in Networked Interactive Virtual Reality. In that paper, I had begun the initial work testing deep leaning as a solution to user hand motion prediction in VR. Since then I have added the ability to train the artificial neural network (ANN) live as the user is interacting to fine-tune the prediction; for such things as individual user motion patterns and fluctuating latency.
The application of our network-based VR is to create a virtual environment where researchers can run interactive coarse-grain simulations, or brainstorm for new ideas while manipulating nano-scale structures with their hands in VR.
Gutmann, G., & Konagaya, A. (2020). Real-time Inferencing and Training of Artificial Neural Network for Adaptive Latency Negation in Distributed Virtual Environments. 2020 International Congress on Human-Computer Interaction, Optimization and Robotic Applications (HORA). doi:10.1109/hora49412.2020.9152833
More detailed information can be found in the links below. Google Translate may be needed.
Information in Japanese can be found here: https://www.titech.ac.jp/news/2020/046181.html
Nano Tech website: https://www.nanotechexpo.jp/
Credit for the work on the Networked VR goes to:
Greg Gutmann: Live trained ANN for motion prediction, simulation system, VR, haptic integration, network protocol
Akihiko Konagaya: Organization of the NEDO project, suggestions/guidance during the research & development
Ryuzo Azuma: DNA origami structure
Fuji Xerox: Development of haptic hardware
Tokyo Institute of Technology: Research environment
NEDO: Artificial Molecular Muscle Project funding
Published: February 2nd 2020