It’s a fun challenge: teaching a quadrupedal robot to dribble a soccer ball well. This is essentially a core component of RoboCup, a large international competition founded in 1996. Football is a great way to test a robot’s locomotion, agility, and decision-making.
There are two main differences from MIT’s Dribblebot. First, RoboCup robots are generally bipedal. Second, and more importantly, the robot is designed to perform complex tasks on bumpy and changing terrain, adding yet another level of difficulty to the task.
In a post related to the news, Yandong Ji, co-leader of the project, said, “Past approaches assumed to simplify the problem of dribbling and model flat, hard ground. Motion.” The robot is also designed to be more static, the robot isn’t trying to run and manipulate the ball at the same time, and this is where the more difficult dynamics come into control. We approached this by extending recent advances that have enabled better locomotion in this complex task, which combines aspects of locomotion and dexterity.”
Potential terrain includes grass (natural), sand, gravel, mud, and snow. The answer to all of the above is one that anyone familiar with the field of robotics these days should know: simulation, simulation, simulation. In training, the physical robot is considered a “digital twin” and puts its paces to the test while the computer runs 4,000 simultaneous simulations of her in different environments.
This kind of training obviously has broader applications beyond the narrow world of robotic soccer. Debate over the effectiveness of legged robots rages on, but one thing is certain. That is, there is a limit to the distance that can be traveled by wheels at present.
“Today, when you look around, most robots are on wheels. But imagine a disaster scenario, flood, or earthquake and robots assisting humans in the search and rescue process. We need machines to navigate terrain, and robots on wheels can’t traverse those terrains,” says MIT professor Pulkit Agrawal. An important part of researching legged robots is to go where current robotic systems cannot reach. ”
Of course, Dribblebot also has its own limitations. Stairs and slopes remain a challenge for small robots.
