Results for ""
Researchers developed DribbleBot, a system for dribbling in the open on various natural terrains, including sand, gravel, mud, and snow, utilising onboard sensing and computing. In the future, such machines may also assist humans in search-and-rescue operations.
Researchers at MIT have made a robot with legs that can dribble a football the same way a person can. The bot used a combination of onboard sensors and computers to move over different types of natural ground, such as sand, gravel, mud, and snow, and to adapt to how each one affected how the ball moved. As a result, "DribbleBot" could get back up and get the ball when it fell, just like any dedicated player.
Programming robot
Programming robots to play football has been a topic of study for a while. But the team wanted the legs to learn how to move when the robot was dribbling automatically. It would let them find hard-to-program skills for surfaces like snow, gravel, sand, grass, and sidewalk. Here comes simulation. The simulation is a computer copy of the real world. It has a robot, a ball, and the ground. From there, it handles the forward simulation of the dynamics. Simulating 4,000 copies of the robot in real-time makes it possible to collect data 4,000 times faster than if only one robot were used. The bot could also negotiate unknown terrain and recover from falls thanks to a recovery controller put into its system by the researchers. This controller allows the robot to continue chasing the ball, assisting it in dealing with disturbances and terrains.
The obsession with robot quadrupeds and football dates back to 1992, when Canadian professor Alan Mackworth gave a presentation titled "On Seeing Robots". Later, Japanese researchers organised a symposium on "Grand Challenges in AI," which sparked debates about how football may be used to promote science and technology. A year later, the initiative was renamed the Robot J-League, and global excitement swiftly ensued. "RoboCup" was born shortly after that.
Experiment
Compared to walking alone, dribbling a football limits DribbleBot's movement and the terrains it can travel. The robot's locomotion must be modified to apply forces on the ball. The ball's engagement with the landscape may differ from the robot's interaction with the landscape, such as dense grass or tarmac. A football, for example, will encounter a drag force on grass that does not exist on pavement, and an incline will impart an acceleration force, altering the ball's steady path. However, as long as the bot doesn't slip, the bot's ability to traverse diverse terrains is frequently less affected by these differences in dynamics. Therefore the soccer test can be sensitive to variations in terrain that locomotion alone isn't.
Conclusion
DribbleBot needs assistance navigating some terrain to make these robots as agile as their natural counterparts. The controller needs to be trained in simulated environments with slopes or staircases. The robot does not perceive the terrain's geometry; it only estimates its material contact properties, such as friction. It is a prospective area of research for the team. The researchers are eager to apply the combined locomotion and object manipulation, such as rapidly transporting various objects from one location to another using the legs or limbs.
Image source: Unsplash
The information provided on this page has been procured through secondary sources. In case you would like to suggest any update, please write to us at support.ai@mail.nasscom.in