Introducing “Atlas”: Pushing the Limits of Robot Mobility
Photo Source: bostondynamics.com
If you’re a fan of animation and vintage TV shows like I am, perhaps you’re familiar with The Jetsons, Hanna-Barbera’s space-age counterpart to The Flintstones. The Jetsons were a “normal” animated family of four, on for three seasons back in the early ‘60’s. There was George, his wife Jane, his daughter Judy, his son Elroy and their maid, Rosie. Four humans and…a robot.
Rosie, the robot maid, was part of “the future.” Well, as NFL head coach George Allen was fond of saying, “the future is now.”
Meet Atlas, an R&D research platform designed to push the limits of whole-body mobility. Atlas’s advanced control system and state-of-the-art hardware give the robot the power and balance to demonstrate human-level agility.
How does that work?
To put it in simple terms, Atlas has one of the world’s most compact mobile hydraulic systems. A custom battery, valves, and a compact hydraulic power unit enable Atlas to deliver high power to any of its 28 hydraulic joints for impressive feats of mobility.
Atlas’s advanced control system enables highly diverse and agile motion, while algorithms reason through complex dynamic interactions between body and environment to plan and execute movements. And 3D printed parts give it the strength-to-weight ratio necessary for leaps and somersaults.
Parkour?
A combination of Asian martial arts, athletics, gymnastics, and obstacle courses, parkour is a training discipline where the goal is to get from one point to another in a complex environment, without assisting equipment and in the fastest and most efficient way possible.
Because it’s a whole-body activity that requires Atlas to maintain its balance in different situations and seamlessly switch between one behavior and another, it gives the Atlas team a perfect sandbox to experiment with new behaviors. By training Atlas to maneuver its way through complex parkour courses, Boston Dynamics engineers develop new movements inspired by human behaviors, extending the limits of humanoid robotics.
Putting Them Through Their “Paces”
In their latest video record of development, the first of the two robots ran up a series of banked plywood panels, broad jumped a gap, and ran up and down stairs in the course set up on the second floor of the Boston Dynamics headquarters. The second robot leapt onto a balance beam and followed the same steps in reverse, and then the first robot vaulted over the beam. Both landed two perfectly synchronized backflips, and the video team captured every move.
But moments after the cameras stop, Boston Dynamics engineers were huddled together, making changes before the next take. Although that attempt was nearly perfect, it was not absolutely perfect. There was one stumble, a slight hesitation. The Atlas team notices every detail and they want to get it right.
Why the Human Form?
A couple of reasons. First, humanoids capture our vision of a go-anywhere, do-anything robot of the future. They may not be the best design for any particular task, but if you wanted to build one platform that could perform a wide variety of physical tasks, we already know that a human form factor is capable of doing that.
Now, a robot’s ability to complete a backflip may never prove useful in a commercial setting, but remember, Atlas is a research platform, not a commercial product. And if robots can eventually respond to their environments with the same level of dexterity as the average adult human, the range of potential applications will be virtually limitless.
The Evolution of Atlas
Five years ago, Atlas was doing the same dive rolls, handstands and backflips it does today, but the underlying processes for controlling those moves have evolved. The previous floor routine and dance moves were about creating a variety of dynamic moves and chaining them together into a routine that could be run over and over again. The robot’s control system still had to make critical adjustments on the fly to maintain balance and posture goals, but the robot was not sensing and reacting to its environment.
Now Atlas’s moves are driven by perception; the robot is adapting behaviors in its repertoire based on what it sees. Engineers don’t need to pre-program jumping motions for all possible platforms and gaps the robot might encounter. Instead, they were able to create fewer template behaviors, all of which can be matched to the environment of the moment and executed online.
Batting .500
Photo Source: bostondynamics.com
In baseball, you’d replace Babe Ruth. In robotics, half the time isn’t good enough. That’s Atlas’s vault average during the filming. When you’re pushing anything to its limit, sometimes that limit is met. Every robot behavior has a small chance of failure. This particular experiment was almost 90 seconds of continuous jumping, jogging, turning, vaulting, and flipping, so those probabilities add up.
The parkour vault in their video isn’t particularly flashy. The obstacle is a balance beam, only a few feet high. Atlas places its arm on the beam and then hoists its body over the structure. For many humans, this sort of vault would be relatively easy, performed without much thought. For the Atlas team, it represented a formidable new challenge.
For one thing, the robot has no spine or shoulder blades, so it doesn’t have the same range of motion that we do. The robot also has a heavy torso and comparatively weak arm joints. Extending their tools to help them find solutions within these constraints was what made the vault an interesting challenge.
The Future
Perfecting it, and other experiments, is what will allow Atlas and its descendants to take us into the future, as depicted (well, somewhat) on the Jetsons. Twenty years from now, I believe that capable mobile robots will move with grace, reliability, and work alongside humans to enrich our lives. But we’re still in the early days of creating that future. Projects like Atlas provide a small glimpse of what’s possible.