Robot chef learns to 'taste' food

Robots are becoming more commonplace in the kitchen, but when it comes to sampling the food they must defer to their human counterparts.

Engineers from the University of Cambridge are hoping to change this with robots that can "taste" as they cook; deciding if a dish needs more seasoning during preparation.

At their lab in Cambridge, the team trained a robotic sensor to detect the salinity—or saltiness—of a plate of scrambled eggs and tomatoes. A small probe on the end of a robotic arm repeatedly samples the food in different places until a comprehensive "taste map" is created by the robot's computer.

To simulate different stages of the chewing process the food was blended into various consistencies. The different readings at different points of "chewing" produced taste maps of each dish.

"The sensor is based on conductance, so it reacts to things like salinity, but it also reacts to things like humidity and consistency of the food. And even though it's just one channel sensing, when we take a lot of samples to get an image of a food, and then we simulate chewing by mechanically processing the food, we can almost get a video of how the food reacts to chewing," explained Grzegorz Sochacki from Cambridge’s Department of Engineering, the study's first author.

While their research is a proof of concept, the team says that by imitating the human processes of chewing and tasting, robots will eventually be able to produce food that humans will enjoy and could be tweaked according to individual tastes.

"We actually don't know much about how the taste perception works," Professor Fumiya Iida, who runs Cambridge’s Bio-Inspired Robotics Laboratory, told Reuters.

"And what we really want to understand... (is) how our taste mechanism works and how we can reverse engineer, so to speak, how we do the tasting."

Published in the journal "Frontiers in Robotics & AI," the "taste as you go" method significantly improved the robot's ability to quickly and accurately assess the saltiness of the dish over other electronic tasting technologies, which only test a single homogenized sample.

The research could help with the development of more advanced automated or semi-automated food preparation by helping robots to learn what tastes good and what doesn't, making them better cooks.

—Reuters