Computer scientists at Princeton are working to bring virtual reality into the physical world, with the potential to enhance a variety of experiences, including remote collaboration, education, entertainment and gaming.
Someday virtual and augmented reality technology will likely be commonplace, said Parastoo Abtahi , assistant professor of computer science. It will be important that users of this technology are able to seamlessly interact with the physical world.
Abtahi and postdoctoral research associate Mohamed Kari are working to make this possible by pairing virtual reality technology with a physical robot that the user can control. Their research will be presented next month at the ACM Symposium on User Interface Software and Technology in Busan, Korea.
Someone using this system can, while wearing a mixed reality headset, select a drink from a list of options and then place it, virtually, on the desk in front of them. Or, more fantastically, they can ask an animated bee to deliver a bag of chips to them on the sofa.
At first, the drink and the chips might be only pixels. But after a minute or so, they will physically materialize, as if by magic. But it's not magic — it's a robot, rendered invisible to the user, that has delivered the snack. "Visually, it feels instantaneous," said Abtahi.
By removing "all unnecessary technical details, even the robot itself," said Kari, the experience appears seamless. The goal is to make the technology disappear and have the interaction between human and computer feel intuitive.
A key technical challenge in this system is communication. The user must be able to communicate their desires simply — selecting a pen across the room, for example, and moving it to the table in front of them. Kari and Abtahi created an interaction technique where a simple hand gesture allows the user to select an object, even from far away.
These gestures are then translated into commands for the robot to execute. The robot is outfitted with its own mixed reality headset, so it knows where to place objects within the virtual environment.
Another technical challenge is adding and erasing physical objects from the user's field of view. Using a technology called 3D Gaussian splatting, Abtahi and Kari create a realistic digital copy of the physical space. Once everything in the room has been scanned, it allows the system to erase something from view, like a moving robot, or add something, like an animated bee.
To achieve this, every inch of the room and every object within it must be scanned and rendered digitally. Right now, the process is somewhat tedious, said Abtahi. Streamlining it, perhaps by assigning the task to a robot, is a subject for future research in her lab.
