An ambitious robot learns to clean a bathroom sink by watching what’s happening.

From cleaning urinals To beach cleaningwe can already see a future in which our robot servants will help make our world a little cleaner. Now a robotic arm has mastered the surprisingly difficult task of cleaning a sink, demonstrating its ability to learn.

Cleaning your sink may not seem like the most difficult task, but when you think about it, a lot goes into it. You must intuitively know at what angle to use the sponge, understand how much force to apply to different parts of the sink depending on the dirt, and constantly adjust your body as you move across the surface. Of course, this is easy for us humans, but if you are a programmer working with a novice robot, you will have to do a lot of coding.

“It is relatively easy to capture the geometric shape of a washbasin using cameras,” says Andreas Kugi from the Institute of Automation and Control at the Technical University of Vienna in Austria. “But this is not a decisive step. It is much more difficult to teach a robot: what type of movement is required for what part of the surface? How fast should the movement be? What is the appropriate angle? What is the right amount of force? “

Realizing that programming all these data points and combinations was a herculean task, Kugi and his team decided to let their robotic arm learn how to perform the task by watching someone else do it.

So they developed a special cleaning sponge, equipped with force and position sensors, and had a person use it to repeatedly clean just the front edge of a sink that had been sprayed with a colored gel that simulated dirt. They then used the data collected from these exercises to train a neural network that could transform the input data into predefined movement patterns. They transferred these patterns to the robot and allowed it to report its movements as it began performing the task. As you can see in the following video, the training went quite well.

Federated learning

Although the experiment focused on cleaning a sink, the researchers say it demonstrates the fact that robotic arms can perform a range of tasks on different surfaces, including sanding, painting or welding sheet metal. What’s more, they say, a fleet of robots could learn basic movements from each other through what they call “federated learning,” and they could then apply those movements to their own individual, specific tasks.

“Let’s imagine that many workshops use these self-learning robots to sand or paint surfaces,” says Kugi. “Then you could let the robots individually gain experience with local data. However, all robots could share the learned parameters with each other.”

Can you say “singularity”?

Document describing the team’s work available at TU Wien. It was recently presented in IROS Conference 2024 and was awarded the “Best Applied Paper” award, distinguishing it from more than 3,500 other papers.

Source: TU Vienna