Weak, Brainless Quadruped Robot Autonomously Generates Gaits

This robot has no sensors, no controller, and weak actuators, but it can autonomously generate a variety of gaits

2 min read
This robot has no sensors, no controller, and weak actuators, but it can autonomously generate a variety of gaits
This robot has no sensors, no controller, and weak actuators, but it can autonomously generate a variety of gaits.
Photo: Osaka University

Roboticists working on quadrupedal locomotion usually spend a lot of time developing control strategies to make their robots more robust and adaptable. The idea is that an advanced controller will let your robot do things on its own more effectively, such as choosing the proper gait for a given task or terrain.

Researchers from Osaka University in Japan are experimenting with a different approach—relying on interactions between the body of a quadruped and its environment to generate gaits without any sensors or controllers or, really, much of anything besides some deliberately weak leg motors.

The most remarkable point of this concept is that the mechanical passivity of a low-torque DC motor behaves as an oscillator model to generate gaits. The low-torque motor in each leg delays and adjusts the phases of the legs by exploiting only a purely physical mechanism: its own weakness.

In these experiments, the only thing that’s being changed is the input voltage to the four DC motors driving the robot’s legs. The gaits that the robot generates arise spontaneously from the interaction between the ground and the robot’s motors: The motors are weak enough that they slow down a bit when there’s a lot of force pushing on them, causing the robot’s limbs to converge on a type of movement where the motors are all synced up with each other. The patterns in which the motors sync depend on their speed, with different speeds resulting in different gaits.

In the experiments, the only thing that’s being changed is the input voltage to the four DC motors driving the robot’s legs. The gaits that the robot generates arise spontaneously from the interaction between the ground and the legs.

The current configuration of the robot exhibited two stable gaits, at 2.5 volts and 4 volts of input. The 2.5-V input resulted in a diagonal-sequence walk, while the 4-V input yielded a transverse gallop. At 1.5 V the robot transitioned between a few different unstable gaits, and 6 V resulted in, well, you saw in the video.

The researchers think that a bunch of this gait variability is inherent to the structure of the robot, including how much its spine flexed in roll and yaw, how much the body segments weighed, and how high the center of gravity was. Essentially, the gaits generated by configuration of the robot are in some sense “natural oscillations,” and a different configuration might “oscillate” differently, leading to different gaits. The researchers also suggest that their experiments may “explain the physical aspects of the gait adjustment mechanism for quadrupeds in nature,” although that particular assertion seems like it could use a bit more evidence to back it up.

And before you say it, we know that whether this is a robot or not is probably debatable—it may technically be more accurate to call it a mechanical automaton. Its creators call it a robot, though, and the paper is being presented at the 2017 IEEE International Conference on Robotics and Biomimetics, so we’re just going to go with a close enough.

“Weak Actuators Generate Adaptive Animal Gaits without a Brain,” by Yoichi Masuda, Keisuke Naniwa, Masato Ishikawa, and Koichi Osuka from Osaka University in Japan, will be presented at IEEE ROBIO 2017 in Macau.

The Conversation (0)

How the U.S. Army Is Turning Robots Into Team Players

Engineers battle the limits of deep learning for battlefield bots

11 min read
Robot with threads near a fallen branch

RoMan, the Army Research Laboratory's robotic manipulator, considers the best way to grasp and move a tree branch at the Adelphi Laboratory Center, in Maryland.

Evan Ackerman
LightGreen

“I should probably not be standing this close," I think to myself, as the robot slowly approaches a large tree branch on the floor in front of me. It's not the size of the branch that makes me nervous—it's that the robot is operating autonomously, and that while I know what it's supposed to do, I'm not entirely sure what it will do. If everything works the way the roboticists at the U.S. Army Research Laboratory (ARL) in Adelphi, Md., expect, the robot will identify the branch, grasp it, and drag it out of the way. These folks know what they're doing, but I've spent enough time around robots that I take a small step backwards anyway.

This article is part of our special report on AI, “The Great AI Reckoning.”

The robot, named RoMan, for Robotic Manipulator, is about the size of a large lawn mower, with a tracked base that helps it handle most kinds of terrain. At the front, it has a squat torso equipped with cameras and depth sensors, as well as a pair of arms that were harvested from a prototype disaster-response robot originally developed at NASA's Jet Propulsion Laboratory for a DARPA robotics competition. RoMan's job today is roadway clearing, a multistep task that ARL wants the robot to complete as autonomously as possible. Instead of instructing the robot to grasp specific objects in specific ways and move them to specific places, the operators tell RoMan to "go clear a path." It's then up to the robot to make all the decisions necessary to achieve that objective.

Keep Reading ↓ Show less