A team of researchers developed DribbleBot, a system designed for real-world dribbling across various natural terrains, including sand, gravel, mud, and snow. 

This device incorporates an onboard sensor and computing capabilities. Furthermore, apart from these notable accomplishments in the field of football, these robotic entities may assist human beings in search-and-rescue operations in the future. It's a familiar sensation if you've ever played soccer with a robot. The sun shines down on your face, and the fragrance of grass fills the air. You take a look around. A four-legged robot is rushing toward you, dribbling with zeal. 

Legged robotic device

Researchers have created a legged robotic device to dribble a soccer ball as humans do. The bot traversed several natural terrains, such as sand, gravel, mud, and snow, using a combination of onboard sensing and computing to adapt to their varied impact on the ball's mobility. "DribbleBot" could get up and recover the ball after falling, just like any dedicated athlete. 

In addition, researchers have been working on programming robots to play soccer for a long time. On the other hand, the team desired to automatically learn how to activate the legs during dribbling to uncover difficult-to-script talents for responding to varied terrains like snow, gravel, sand, grass, and concrete. Here comes simulation. 

Simulation

Inside the simulation - a digital replica of the natural world — are a robot, a ball, and terrain. Load in the bot and other assets, establish the physics parameters, and it will handle the forward simulation of the dynamics from there. Four thousand robot versions are simulated in parallel in real-time, allowing data collection to be 4,000 times faster than with a single robot. That's a lot of information. 

The bot's recovery controller allowed it to navigate obstacles and get up after being knocked over. This controller helps the robot recover from out-of-distribution interruptions and terrains, allowing it to continue dribbling after a fall and following the ball. 

DribbleBot

Compared to walking alone, dribbling a soccer ball limits DribbleBot's movement and the terrains it can travel. A soccer ball, 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 regular course. 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. 

The robot contains sensors that allow it to detect its surroundings, feel where it is, "understand" its position, and "see" part of its surroundings. It has a collection of actuators to apply forces and move itself and things. The computer, or "brain," stands between the sensors and actuators, tasked with processing sensor input into actions that will be applied via the motors. When the robot runs on snow, it cannot see but feel it via its motor sensors. However, soccer is a more difficult accomplishment than walking, so the scientists used cameras on the robot's head and body to add a new sensory modality of vision to the new physical skill. After that, we dribble. 

Conclusion

DribbleBot struggled on several surfaces because there is still a long way to go before these robots are as agile as their natural counterparts. There are no plans to train the controller on simulated terrains, including hills and stairs. The robot merely makes educated guesses about the friction and other material contact qualities of the ground it's moving across. In the future, the researchers would like to investigate what happens when the robot encounters an obstacle it can't overcome, such as a step up. The team behind DribbleBot is eager to take what they learnt while building the robot and apply it to future projects that require robots to move and manipulate objects simultaneously.

Image source: Unsplash