Teaching a four-legged robot to dribble a football successfully is an interesting problem to solve. It’s basically necessary to RoboCup, the massive international tournament that began in 1996. The skills of mobility, agility, and decision making are all put to the test in the game of football.
Here are two major distinctions from MIT’s Dribblebot: To begin, most of the RoboCup robots walk on two legs. Second, and most critically, itis built to carry out the intricate duty over uneven and changeable terrain, which significantly increases the complexity of already challenging endeavour.
‘Previous approaches simplify the dribbling problem, making a modelling assumption of flat, firm ground,’ writes Yandong Ji, one of the project’s co-leaders, in a blog post about the development. Because the robot isn’t attempting to run while manipulating the ball, the motion is also planned to be more static. The control problem becomes more challenging when dynamical complexity is introduced. We overcame this challenge by using previous developments that improved outdoor mobility to this hybrid task that required both walking and dexterous manipulation.
Grass (of course), sand, gravel, mud, snow, and snowy mud are all possible surfaces. Anyone who has even a passing knowledge with the robotics space as of late should know the answer to all of the above: simulation, simulation, simulation. The actual device, which serves as a “digital twin” in the training process, is put through its paces while four thousand virtual surroundings are simulated on computers at once.
The robot has utility beside playing as well
Obviously, this type of training may be used for more than only robot football. The effectiveness of legged machines is hotly contested, but one thing is certain: you can only get so far on wheels right now.
“Most robots today use wheels for mobility. So let’s say water or an earthquake strikes and we need them to help people throughout the search and rescue operation. Wheeled robots can’t traverse those environments,” says MIT scientist Pulkit Agrawal, “and we need the machines to navigate over terrains that aren’t flat. Researching legged robots is important because they can explore environments that conventional can’t.
Dribblebot has its own shortcomings, of course. The little machine has some work to do before it can conquer stairs and slopes.
To read our article about “With PMI at 58.7, Saudi non-oil economy is still booming” click here.
