MIT researchers show use of muscle signals to control drones

For novices of drones, different styles of joysticks take longer to adapt and become skilled. But researchers at MIT have come up with a new way to use the operator’s muscle signals to more intuitively control drones. Equipped with multiple electromyogram sensors, the Conduct-A-Bot system is able to wear the user’s right arm biceps, triceps and forearms areas to read the user’s muscle signals.

MIT researchers show use of muscle signals to control drones

Joseph Del Pretoto demonstrates muscle control and guides the drone through the iron ring (from: MIT)

The companion sensors detect muscle and arm activity and relay data to a microprocessor with which it is hard-wired to identify different arm movements with machine learning-based algorithms.

Each action has been preprogrammed to easily convert the user’s action to a specific instruction and then wirelessly transmitted to the Parrot Bebop 2, a quadcopter.

MIT researchers show use of muscle signals to control drones

The system can also be fine-tuned and adapted to each user’s unique iemosignal signal (from:MIT)

By default, the taut upper arm allows the drone to hover, clenching the fist means moving forward, turning clockwise/counterclockwise, and swinging up and down (horizontal shift).

During the most recent test, Bebop correctly responded to 82% of the 1500 commands. It is believed that this number will be further increased as the system develops further.

Controlling Drone with Gestures (via)

“This system is an important step forward for seamless collaboration between humans and machines, making it a more effective and intelligent tool for everyday tasks,” says Joseph DelPreto, a researcher.

Looking ahead, the technology will not only be compatible with other drones, but can even be used as a secondary robot to help older people or people with disabilities live a more comfortable life.