Frustration and The pain of trying to piece together IKEA furniture and failing to piece together may seem like a humiliation to you, but be aware: the nightmare of particleboard may one day cause robots to become less stupid.
In recent years, robotics experts have discovered that building IKEA furniture is actually a good way to teach robots how to deal with the chaos of the real world.A group of researchers wrote a simulator in which the virtual robotic arm uses trial and error Put the chairs together. However, others managed to obtain a different set of robotic arms to build IKEA chairs in the real world They spent 20 minutesNow, a useful robot can help humans assemble Ikea bookcases by predicting what parts they want next and handing them over.
Stefanos Nikolaidis, a roboticist at the University of Southern California, said: “This is one of the things that is easy to try-even if we break a few bookcases in the lab, it’s not a big deal.” Recent papers Describes this research, which was published at the International Conference on Robotics and Automation in May. “It’s cheap. It’s also something we have to do at some point in our lives.”
Nikolaidis and his colleagues began to study how different people built IKEA bookcases. They did not provide them with instructions with hieroglyphs, but asked the subjects to improvise the order in which they arranged support plates and shelf inserts for the frame. (This is an important distinction, because the larger research question of this experiment is not about building furniture—more on that later.) Based on these results, researchers can classify people into types or preferences. For example, some people connect all the shelves to one of the frames. Others will connect a shelf to two frames at the same time. These are called action sequences.
Then they asked the subjects to assemble it again, this time with a robotic arm nearby grabbing debris for them. The researcher will record which part (shelves or supports) the person starts from, and build a clue for the robot. “Suppose you come in and place the first shelf,” Nikoladis said. “Well, the robot doesn’t know much. Then you choose the second shelf. Now you start to place the third shelf. Well, you are likely to belong to the group of users who assembled all six shelves in a row. This is very very Not too possible Then you will suddenly change your preferences. “Once a robot knows a person’s preferences, it will give them the previously selected part of the person it knows. Experiments have shown that the robot can quickly and accurately adapt to the human style in this way and successfully deliver the correct components.
Think of it as the way artificial intelligence researchers develop image recognition algorithms: if you want to detect cats, you can feed a large number of images of cats to the neural network. Because I have seen so many examples before, the algorithm can be generalized. If you show it a photo of a cat it has never seen before, it can use its previous knowledge to confirm that it is indeed parsing a furry four-legged mammal, with a bad attitude.
This robot is doing the same thing, but instead of using a set of static images, it borrows the following example sequence, Humans piece together the order of shelves and brackets according to their own preferences. Nikolaidis said: “The robot knows that the next action it should do is to hand over the next shelf to you, very, very sure.”
However, in the final analysis, this research is not about developing highly specialized robots that can come to your home to help you build a bookcase. Nor is it to develop a machine that can perform such complex tasks on its own.It’s about teaching robots how to cooperate with humans without driving them more Than the craziness people have gotten when building IKEA furniture.
Despite all the hustle and bustle about robots coming to steal our work, the reality is that you are more likely Let the machine work with you ratio Directly replace youFor the time being-maybe for a long time in the future-people will only do better on certain tasks. No machine can replicate the dexterity of a human hand, or get close to solving problems like we do.What robot Yes What I am good at is brute force work. Think about the car assembly line: the robot arm lifts the door into place, but the fine details need to be humanized.