Marlo conquers uneven terrain in sporty shoes. The new robot, Marlo, was created by University of Michigan and Oregon State University. (via University of Michigan)
There have been several innovations in the field of robotics mainly bipedal robots, but many of them still have issues waling unaided. Instead of looking fierce and ready for anything, these bots often look like unsteady toddlers. Boston Dynamics has successfully built robots with great stability and the capability to walk on uneven terrain, such as Atlas, but now a new robot is entering the field and it may be even better.
The University of Michigan's Jessy Grizzle and Oregon State University's Jonathan Hurst have created an unsupported bipedal robot in a joint project between the two schools. The robot, named Marlo, can walk down steep slopes, through thin layers of snow, and over uneven and unstable ground. But there's a key difference to how Marlo achieves this: through algorithms. How it works is it has a bank of algorithms which contains instructions for various walking styles. Marlo analyzes data from sensors in her knees, hips, and torso to adjust the walking style on the fly.
"The robot has no feeling in her tiny feet, but she senses the angles of her joints—for instance, her knee angles, hip angles and the rotation angle of her torso," said Jessy Grizzle, professor of electrical engineering and computer science and of mechanical engineering. "It's like walking blindfolded and on stilts."
This makes Marlo Grizzle's first robot that can walk and fall in any direction, which is known as 3D walking. Grizzle's team previously produced Mabel was attached to a boo, giving her sideways stability. Mabel also allowed the team to begin working on the algorithms they would eventually use for Marlo. The team has taken what they learned from Mable and created a way to control the robot using 2D algorithms.
The robot's speed and direction is determined by someone holding an Xbox controller (see, video games are useful). Other factors like movement speed is completely handled by the robot herself albeit blindly. The robot adjusts her gait by determining the changes in ground height and terrain. It sounds pretty impressive, but that doesn't mean the robot doesn't have its limitations. Marlo works as long as it doesn't have to make quick turns or sideways movements. This issue can only be resolved by creating a more integrated algorithm. Grizzle's team is already working to resolve this by developing a fully 3D controller.
Though the team hopes their findings can be used to improve all bipedal robots, they also believe the technology can be extended to robotic prosthetics. "The ability of MARLO to gracefully navigate uneven terrains is very exciting for my work in prosthetics," said Robert Gregg, an assistant professor of mechanical engineering and bioengineering at the University of Texas-Dallas. "We hope to encode similar abilities into our robotic prosthetic leg so that lower-limb amputees can just as easily walk about the community without having to think about the terrain."
If this technology can be implemented in prosthetics, then it's even more useful than we think. But it doesn't change the fact that Marlo looks like something out of a Terminator movie. So while we marvel at the way Marlo walks through snow, let's keep an eye on her to make sure she doesn't turn against her creators Ex Mechnia style.
Have a story tip? Message me at: