A long farewell to Baxter, the gentlest giant among robots

For a serious research robot, Baxter looks charming. Red color inherent in sports cars, two large and careful hands-manipulators. His face is a flat screen that conveys “feelings” of embarrassment (red cheeks, raised eyebrows). Those interested can sit in front of him and ask him to read their thoughts, correcting the mistakes. Or you can point him to objects that should be raised. If he gets confused, he can ask you a clarifying question - this seemingly simple interaction is in fact a big breakthrough for a promising area of ​​human and robot communication.

Similar studies have made Baxter a legend among engineers.at the beginning of our era of advanced robotics. Machines are no longer limited to factories - they begin to roll and walk beside us. So, we need to understand how we, hell communicate with them - and them with us. Baxter is the most important representative of the avant-garde, engaged in this issue, providing researchers with a platform for solving such serious problems as object manipulation.

But last week, the manufacturer of this machine, Rethink Robotics , suddenly announced that it plans to close the shop, as sales of robots have plummeted. (Baxter was originally developed for use in production, for example, for lifting objects and placing them on a conveyor belt, and then adapted for research). The announcement came as a surprise to everyone, but probably not so unexpected - the industry is overwhelmed with a booming mass of competitors seeking to occupy the market.

Baxter is not the only machine available to robotics, so research on the issues of communication between humans and robots will not arise. However, Baxter, as well as his next-handed generation, Sawyer, indispensable for research in the field of robotics. They may not have crushed the entire industry under themselves, but they definitely staked out their place as pioneers in the region.

“All robotics researchers today either have their own Baxter, or have a friend with Baxter,” says a robotics specialist from Brown University, Stephanie Telex, herself in the first group. "In terms of penetration, it is closest to a common manipulation platform for various research laboratories."

Baxter became special because it was an all-in-one solution — a camera, hands, clamps, and sensors. It was released in 2012, and it became an independent creature that robotics could program and change at will. In addition, it was extremely cheap, at least by the standards of robots - Telex bought Baxter for only $ 27,000, while other robotic platforms could cost you hundreds of thousands of dollars.

But this was a compromise. In general, there are two approaches to creating a robotic arm. The rude robot working on the assembly line acts with high precision as necessary - otherwise, your cars will have the doors bolted to the roof. Accordingly, its drive mechanisms (electric motors that control the hand) are very expensive.

Another approach is to abandon extreme precision in favor of sensations. This is how collaborative robots , or cobots, work. If they work with people instead of car parts, you are more concerned that they do not smash someone's skull than their 100% accuracy. Therefore, although Baxter is not a particularly accurate robot, its sensors allow it to detect the appearance of contact with a person and stop in time. This makes it ideal for researchers who do not want their robots to hurl them throughout the laboratory.

“I wanted to let go, because he had two hands and it was safe to work with him,” says Rodney Brooks, founder of Rethink. “It was not necessary to close access to it for undergraduates. It was possible to leave students with him for the whole night, and they work precisely at night, and no one was injured. ”

But it is not necessary to trust Brooks with the word: “This was one of his properties, which allowed us to give him younger students and not worry too much about it,” Telex says.

So, students could experiment with the manipulative capabilities of the robot without fear of complications. But researchers also began working with Baxter as part of research on the interaction of robots and humans (human-robot interaction, HRI). And here, at least from my point of view, includes the establishment of some kind of invisible connection between man and machine. She has a flat screen on her face.

“And everyone immediately started placing their faces on the screens,” says Brooks. - And many researchers engaged in HRI. I did not plan this, I thought only about the study of manipulation. "

But it cannot be said that object manipulation and HRI cannot go hand in hand (sorry for the pun). Telex in its work uses Baxter to study the interaction of people and machines in unexpected and unexplored ways. For example, her robot can ask a clarifying question, if you are not sure what a person needs to raise.

And let's not forget that, although Rethink did not sell as many robots as it wanted, Baxter nevertheless became a serious step away from the bulky insensitive industrial robots that existed before it. It was easier and safer. “Baxter proved that industrial robots can work safely with people in the workshop,” said Rick Falk, director of warehouse robots at Locus Robotics. "More importantly, he demonstrated that effective programming of robots does not have to be difficult, and this allows you to perform a wider range of activities in collaboration robotics."

Baxter and his relative Sawyer entered a new and unknown territory. American companies are turning to cobos to keep up with competitors. And the competition in the field of collaboration between humans and robots is becoming fiercer. “At Rethink Robotics, we imagined something inspiring and practical at the same time: to give safe robots that are easy to assign with the help of intuitive interfaces into the hands of any factory worker,” says Daniela Ras, director of the Laboratory for Computer Science and Artificial Intelligence at MIT. “Baxter and Sawyer robots urged the robotics community to push the boundaries of robotic manipulation capabilities and helped small and medium-sized companies organize their own automation.”

However, this does not mean that tomorrow robotics will find that the Baxters have disappeared from their laboratories. These machines will continue to work as pioneering research platforms. Rethink Robotics is closing, but Baxter will continue to live until, inevitably, something better appears.

“We are ready for a new platform,” says Telex. - We are ready for something else. And that's fine. Part of the advancement of our field depends on trying to test our technology on new, different from old robots, and to understand what common problems exist and how they can be solved for all robots. ”

Goodbye, Baxter. You did a good job.