Innovators Update Season 2: Ep 4: Using muscle signals to control robots

author avatar

10 Jun, 2021

Joseph DelPreto is a PhD candidate at MIT CSAIL working on making human-robot interfaces more accessible and effective by using wearable sensors and biosignals.

Joseph DelPreto is a PhD candidate at MIT CSAIL working on making human-robot interfaces more accessible and effective by using wearable sensors and biosignals.  

DelPreto’s research explores how using muscle, motion, and brain signals could infer how a person wants a robot to behave  during physical tasks without external sensing infrastructure or tedious setup. Developing user-independent classification pipelines that do not require extensive calibration or training data could allow a new user to simply don an unobtrusive wearable sensor and begin interacting with a robot.

Two systems that explore these concepts are RoboRaise and Conduct-a-Bot.  RoboRaise focuses on direct physical collaboration - it allows a robot to help a person lift objects by processing biceps and triceps muscle signals. Conduct-a-Bot focuses on remote control of mobile robots - it allows a person to teleoperate a drone by making gestures, which are detected using muscle and motion signals from the forearm and upper arm. These systems can hopefully move towards robots that fluidly help people perform assemblies in industrial settings, move furniture around a house, survey large areas, or perform tasks in dangerous locations.


“When we think of a future with ubiquitous robots, we think of robots in homes, factories, and remote areas extending our capabilities and seamlessly providing assistance.  Yet at the moment, there is often a communication barrier when trying to interact with machines,” says DelPreto. “Advancing towards the vision of readily available robot teammates for all users will require intuitive ways of telling robots what we want them to do without any tedious training or setup - a way for the robot to adapt to the person rather than the other way around.” 


Prior to his PhD research, DelPreto received a Master of Science degree from MIT in Electrical Engineering and Computer Science, with a minor in Finance and Business.  He received a Bachelor of Science degree from Columbia University in Electrical Engineering, with minors in Mechanical Engineering and Psychology.

This episode is part of our Innovators Update with Rachel Gordon at MIT. Check out previous episodes here.

This video was produced by Rachel Gordon from MIT CSAIL and Richard Hulskes from Wevolver.   






More by Rachel Gordon

Communications and Media Relations Manager at CSAIL, MIT’s Computer Science and Artificial Intelligence Laboratory pioneers research in computing that improves the way people work, play, and learn.

Wevolver 2022