Docket #: S19-463
RoboTurk: A Cloud-based Crowdsourcing Framework for Real-Time 6-DoF Robot Control with Handheld Mobile Devices
Stanford researchers have developed a crowdsourced framework for real-time robotic teleoperation with six degrees of freedom. Through smartphone controllers, RoboTurk enables large human workforces to remotely operate the robots without the need for prior training. The use of handheld mobile devices allows for natural, intuitive control of the robot. Data collected through the framework can further be used to train robots to complete other autonomous tasks.
Stage of Research
Applications
- On-demand human workforce for robotic applications
- Large-scale remote data collection
- Can be used to train autonomous robots
Advantages
- No training from experts on robot operation
- 6 Degrees of freedom control without special hardware
- Only requires a smartphone and web browser
- Real-time teleoperation via large, remote workforces on demand
Publications
- Mandlekar et al. arXiv (2019) " RoboTurk: A crowdsourcing platform for robotic skill learning through imitation"
Marketing
"Robots learn tasks from people with framework developed by Stanford researchers", October 26, 2018
Related Links
Patents
- Published Application: 20230226696
Similar Technologies
-
StickyBot - Climbing with dry adhesives S06-091StickyBot - Climbing with dry adhesives
-
Ocean One: Robotic Avatar for Extending Human Reach S16-446Ocean One: Robotic Avatar for Extending Human Reach
-
Manufacturing Method for Synthetic Gecko-Inspired Adhesives S10-333Manufacturing Method for Synthetic Gecko-Inspired Adhesives