“This vision of designing robots as intelligent tools of the future, of creating machines that can help us with cognitive tasks as well as physical tasks, is ultimately what drew me to MIT and CSAIL,” says PhD student Joseph DelPreto, whose research focuses on human-robot interaction. Specifically, his goal is creating robots that are accessible, effective teammates.
A member of Daniela Rus’ Distributed Robotics Lab, DelPreto imagines a future where we turn to robots for assistance with physical tasks in the same manner that we currently turn to apps when confronted with computational tasks. “To work towards this vision, I’ve been exploring the use of muscle and brain signals to create more natural ways of physically collaborating with robots, supervising robots, and teaching robots new tasks,” he shares.
DelPreto has worked to leverage wearable sensors and biosignals to enhance communication between people and robots: “We’ve created machine learning pipelines and signal-processing algorithms that can use brain and muscle signals for communication.” The wearable surface EMG electrodes collect biceps and triceps activity, and, as the robot processes this information, it is able to mimic a person's motions or detect nonverbal cues to lift an object with a person the same way two people would lift something together.
According to DelPreto, this technology can be utilized in a variety of contexts. “Imagine a construction zone or a factory floor, where a worker needs to install some insulation on a wall or ceiling. Or imagine you finally have a robotic assistant in your house, and you want to hang a poster on the wall together,” DelPreto explains, “In all of these cases, a robot assistant that could help lift and manipulate a large or heavy object just by processing your natural muscle signals could be very helpful.”
The ability of humans and robots to interact in an intuitive manner could increase both productivity and quality of life by leveraging the respective strengths of humans and machines. “For example, if robots can focus on dull, dirty, or dangerous tasks, then people can focus more on strategic, creative, or emotional tasks,” says DelPreto. He envisions a future where robots truly complement humans, and this synergy could provide a tremendous amount of computational and physical support.