In many ways, the future that we live in today offers a lot more than what was promised in the past. Your smartphone alone is an almost magical piece of technology that few people could have predicted. But in other ways, this future we’re living in is a disappointment. We don’t have flying cars, we don’t have a thriving lunar colony, and we certainly don’t have the kind of robot servants that were promised in fiction like The Jetsons. Flying cars are a pretty terrible idea to begin with and there is little practical reason to build a lunar colony, but a team of roboticists have at least built a robot that can learn how to make the ideal omelet for you.
OmeletteBot was created by a team from the University of Cambridge and presented at the ICRA 2020 conference in Paris this past week. Omelets really aren’t the point of the project, which is really just an exploration of how automated systems can alter their approach to suit the needs of each individual. The omelet was chosen over other potential dishes because it’s easy enough for a robot to make, while still providing enough flexibility that its parameters can be tweaked according to individual preferences.
The robot itself is a pretty standard arm equipped with a gripper end effector that can be used to pick up ingredients, the mixing bowl, and the pan. The parameters that can be adjusted are the amount of salt, amount of pepper, the time spent whisking, the time spent mixing in the pan, and the time that the pan is left on the heat. That’s a pretty basic omelet, so don’t expect anything fancy like peppers or ham. An individual can try samples of a few omelets that have minor differences and then rate those based on flavor, appearance, and texture. OmeletteBot can then use a process called batch Bayesian optimization to determine what parameters will result in the ideal omelet for that individual and then cook it. It’s unlikely that your local diner will have robotic line cooks anytime soon, but OmeletteBot is an interesting look at how robots can be taught to serve us better.