Researchers from the University of Cincinnati are teaching autonomous robots to open doors and locate wall outlets using machine learning.

Opening doors has consistently been a challenge for robots due to doors' different shapes, sizes, handles and forces required to open them, among other factors. Previous efforts involved mapping and creating 3D digital models of the room so that the robot can locate doors within that room specifically. However, that approach is time-consuming and only works for the room that has been scanned.

As such, the team from the University of Cincinnati employed machine learning to “teach” the robot how to open doors via trial and error.

"The robot needs sufficient data or 'experiences' to help train it," explained the researchers. "This is a big challenge for other robotic applications using AI-based approaches for accomplishing real-world tasks."

For more on the development, watch the accompanying video that appears courtesy of the University of Cincinnati.

The study, Force-Vision Sensor Fusion Improves Learning-Based Approach for Self-Closing Door Pulling, appears in the journal IEEE Access.

To contact the author of this article, email