DARPA Robotics Challenge recently ended, but the massive event that inspires innovation saw more than 20 teams from all corners of the world come together in order to celebrate advancements in the field of robotics and explore how robots can improve the daily lives of humans, maybe someday even save a life.
The very goal of the challenge was to inspire robot engineers to design robots that are able to perform search-and-rescue missions. The idea behind the theme is that certain disaster areas are too dangerous for human rescuers to enter.
Gill Pratt, DARPA program manager and DRC organizer, gave a statement during the award ceremony saying that “These robots are big and made of lots of metal and you might assume people seeing them would be filled with fear and anxiety. But we heard groans of sympathy when those robots fell. And what did people do every time a robot scored a point? They cheered”.
Despite what one could initially expect from such an event, the focus was not on developing autonomous robots that could perform tasks all on their own, without the assistance of a human controller. The noticeable absence of autonomous robots was due to the great level of difficulty in creating and operating such a machine.
Instead robot designers had to remotely guide their creations through eight (8) different tests, in the simulation of a disaster zone. Not all robots managed to get through all eight (8) tests, for some they were too difficult.
Participating teams had to either build or program a robot before controlling them remotely and showing off how well they can perform tasks such as drive a car, climb stairs, open a door, cut a hole in the wall and cross over rubble, all in under one hour.
The winners, a team of robot engineers from the Korean university KAIST, got their robot to finish all eight (8) tasks in less than 45 minutes and received $2 million from DARPA in return. The robot itself is named DRC-Hubo.
The time of the winning robot is still not as good as hoped, which shows that while the tasks were simple enough for a human to do, it turns out that robots needed more time.
There are two (2) kinds of control in robotics. One is low-level control which coordinates the actions of single, independent motors. The other one is high-level control which carries out tasks using the entire system. Both high-level control and some of the low-level control were carried out by human operators during the challenge.
Even though they weren’t limited to it by the rules, most of the participants chose to design machines that were similar to humans in their appearance. This showed how difficult it is to design robots as it implies a great level of coordination. When a human-looking robots executes an action with a part of its body, all the other parts have to be coordinated as well in order to counteract the forces involved.
This was well exemplified when robots had to drill a hole through a wall. They had to generate enough force and energy to actually damage the wall while also having to constantly readjust its balance so that it wouldn’t fall over.
Humanoid robots have a particularly complex shape that could be hiding more than 30 different joints, and all them could be moved simultaneously. What this means is that low-level control was particularly important for DARPA designers.
Experts inform that as the low-level control and quality of the body of human-looking robots improves, scientists could begin to move towards creating true artificial intelligence (AI).
Image Source: livescience.com