MIT engineers have developed a new method designed to make robot training more accessible, reducing the need for specialised programming skills. The approach combines three established “Learning from Demonstration” (LfD) techniques into a single sensor-equipped tool that researchers say could broaden the use of robots in manufacturing, household assistance and caregiving.
The new device, known as the Versatile Demonstration Interface (VDI), integrates teleoperation, kinesthetic training and natural teaching into one handheld module that can be attached to common collaborative robot arms. The tool includes a camera, position-tracking markers and force sensors, allowing robots to learn tasks by observing human demonstrations. MIT said the system “makes training simpler and more flexible,” aiming to open robotics to a wider range of users.
See also: MIT Researchers Create Energy-Storing Concrete for Future Infrastructure
Researchers tested the VDI with volunteers performing tasks such as press-fitting and moulding. According to the team, participants with manufacturing experience often selected the natural teaching method, reflecting familiarity with hands-on workflows. The study indicated that combining multiple LfD modes in one tool may make robot instruction more intuitive across different skill levels.
MIT said the training system can be used with industrial robots, but could also support emerging use cases in domestic and caregiving environments. The researchers highlighted growing interest in household-assistance robots and noted that more adaptable teaching methods could accelerate adoption.
See also: XPeng’s Humanoid Robot Iron Stuns Audience With Lifelike Movements at Live Debut
One of the project’s lead researchers said the goal is to enable robots to work effectively alongside people, adding that flexible demonstration tools could support “domains where we hope to see increased robot adoption.”
Source: MIT