Why is it relevant to do it remotely?

Remote teaching of robots lets experts demonstrate tasks from anywhere, without being physically next to the robot. That’s especially useful when the robot is in a hard-to-reach place—like in surgery, space, or even in a factory on the other side of the world. It also makes it easier to scale expertise: one skilled operator can teach multiple systems without traveling (democratizing access to skills).

What are your main challenges?

The big challenge is the network—delays and data loss can really affect haptic feedback fidelity and the accuracy of the demo. This is even more pronounced when you are performing contact-rich tasks that rely on precise force feedback. To solve this, we’ve built a system that recognizes when parts of a demonstration are affected by delay or noise, and adjusts how much they influence the robot’s learning, improving robustness and autonomy.

Who in industry is looking for this?

This kind of solution is relevant for industries like healthcare, manufacturing, and aerospace—places where precision, safety, and remote access matter. Also, companies building collaborative robots or teleoperation systems are looking for better ways to make remote interaction more intuitive and reliable.