For decades, programming industrial robots required specialized coding skills, creating a barrier to their flexible deployment. MIT researchers are dismantling that barrier with a breakthrough handheld device that enables anyone to train collaborative robots using their preferred teaching method. The "versatile demonstration interface" (VDI) attaches to standard robotic arms, empowering users with three intuitive training approaches:

  1. Teleoperation: Control the robot remotely via a joystick-like interface
  2. Kinesthetic Training: Physically move the robot through desired motions
  3. Natural Teaching: Perform the task yourself while the robot observes and learns

"We are trying to create highly intelligent and skilled teammates that effectively work with humans to get complex work done," explains Mike Hagenow, MIT postdoc and lead researcher. "Flexible demonstration tools can help far beyond the manufacturing floor—in homes or caregiving settings."

How the VDI Works

The sensor-packed attachment includes:
- Motion-tracking cameras/markers capturing spatial positioning
- Force sensors measuring pressure during tasks
- Interchangeable mounting for different robot models

During testing at a manufacturing innovation center, experts used the VDI to teach robots two common tasks:

Task Description Training Method Preferences
Press-fitting Inserting pegs into holes precisely Natural teaching favored for precision
Molding Spreading material evenly on a rod Kinesthetic useful for force adjustment

Users overwhelmingly preferred natural teaching for its intuitiveness but noted distinct advantages for each method:
- Teleoperation for hazardous material handling
- Kinesthetic for adjusting heavy payload positioning
- Natural teaching for delicate, nuanced maneuvers

Beyond the Factory Floor

This flexibility addresses critical adoption barriers:
- Democratization: Non-programmers can now train robots
- Task Diversity: One robot can learn wider skill sets from multiple teachers
- Safety: Remote training minimizes exposure to dangerous processes

As collaborative robots expand into healthcare and homes, MIT's interface provides the missing link: letting humans teach machines their way. The team is refining the design based on user feedback, signaling a future where adaptable robots learn organically from diverse human expertise—no PhD required.

Source: MIT News (July 17, 2025), 'New tool gives anyone ability to train robot'