A group of researchers recently presented new technology that would help robots, namely those used in a home, to better perceive and process 3D objects.

The robot perception algorithm, which was unveiled in July at the 2017 Robotics: Science and Systems Conference in Cambridge, Mass., allows the robot to guess what an object is, how it’s positioned and what it looks like if parts are hidden without looking at the object from multiple angles – mimicking how humans view objects.

Researcher and Duke University graduate student Ben Burchfiel explained this ability is particularly important for robots that would operate within a home where objects and the environment are less organized than in a lab or on a factory floor. Robots would need to be able to perceive 3D objects from one view to carry out household operations like clearing a table.

First, the algorithm receives training in the form of thousands of complete 3D scans of household objects. Then, the algorithm groups the scans into categories based on how similar objects are to each other. When viewing something new, the robot can sift through categories of objects, rather than each object, in its database to determine what the new object probably is or what its hidden parts look like.

“It’s impractical to assume a robot has a detailed 3D model of every possible object it might encounter, in advance,” Burchfiel said in a Eurekalert release.

To test the algorithm, researchers fed it 908 new 3D examples from the established categories, but from the top view to determine how many items the algorithm could correctly guess. The researchers found that the algorithm correctly guessed what a new object was 75% of the time. The competing technology did so only about 50% of the time.

Unlike the other technologies, the new algorithm can also recognize items that are rotated. But it will make mistakes when items resemble the shape of other objects from certain perspectives.

“Overall, we make a mistake a little less than 25% of the time, and the best alternative makes a mistake almost half the time, so it is a big improvement,” Burchfiel said. “But it still isn’t ready to move into your house. You don’t want it putting a pillow in the dishwasher.”