If you find yourself in a Johns Hopkins University laboratory that appears to be overrun by cockroaches, don’t immediately call the exterminator.
A JHU team has developed a prototype robot that steals some moves from a Central American insect species known as blaberus discoidalis. The researchers hope that their motion-control strategy, modeled after the resilient roach, will someday support search-and-rescue applications that call for traversal on a variety of terrains.
Sean W. Gart, a postdoctoral fellow, and Chen Li, director of the university’s Terradynamics Lab, led the study and co-authored two related papers published this month in the journal Bioinspiration & Biomimetics.
A Roach Approach
So, what’s so special about the movements of a two-inch-long cave bug?
The discoid roaches, native to Central and South America, are fast, operate in a variety of environments, and can take a hit, responding well to external perturbations, says Gart.
Or to put it more simply: “They’re just hard to kill.”
In a rainforest environment, the Central American cockroaches contort their heads, torsos, and legs to move across obstacles and remain on course.
“Where they live, you have all sorts of stuff around you, like dense vegetation or fallen leaves or branches or roots," Li said in a press release. "We're trying to understand the principles of how they go through such a complex terrain, and we hope to then transfer those principles to advanced robots."
To create the robot, Li and Gart first had to study the cockroach itself. Inside the lab, the team set up an obstacle track featuring large "bumps" and "gaps" mimicking the cockroach’s natural, vegetation-cluttered habitat.
As the roaches were funneled through the course, high-speed cameras captured the insects' body and leg motions. By slowing the videos down, the researchers learned the precise tactics that small robots might also feature to surmount the same types of obstacles.
“We came up with a generalized model that we can use to predict how large of a gap that can be traversed, given your speed, your initial body angle, or pitch,” Gart told Tech Briefs.
Li's team then built a six-legged robot to replicate the insect's running patterns; with the push of a button, the feet-forward robot runs toward a given obstacle.
The legs operate in a bio-inspired alternating tripod gait: three legs touch the ground at all times, switching regularly from one side of the body to the other.
Along with a roach-like rounded shell, the researchers added a "tail" to give the robots a greater ability to overcome bumps and gaps. The tail swings up and backwards, transferring angular momentum to the body and allowing the six-legged system to pitch upwards prior to encountering an obstacle.
Through the study, the researcher learned that the simple control strategy of actively pitching the robot up in the air increases the performance of obstacle traversal over a gap or bump – without the need for complicated maneuvers or sensors.
Pitching supports object traversal because the robot can lift its front end up and over the bump edge, instead of directly colliding with it. Additionally, a pitched-up body increases height, thus increasing the gap length that can be bridged.
According to the JHU team, the tail allowed the robot to increase gap traversals by 50 percent and bump traversals by 75 percent.
The researchers found that the cockroach, even if it was barely able to clear the gap or climb the bump, was often able to climb out – thanks to a great grip.
“If our robot just barely crossed the gap, it still fell in more often, just because the cockroach has such specialized feet for gripping surfaces,” said Gart.
One area for the researchers to take the study would be to recreate the cockroach’s hold capabilities, said the JHU post-doc.
For now, the Johns Hopkins team is testing the system on new terrains, including grass stalks and pillars that are more flexible.
By studying the fundamental principles of object traversal, the team hopes to apply the idea to search-and-rescue robots – technologies that may have to operate through rubble after, say, a landslide or a hurricane.
“We noticed that these types of terrains are fundamentally made of gaps and bumps,” Gart told Tech Briefs. “That‘s why we chose to study these types of obstacles.”
By developing low-level control strategy for roach-like obstacle avoidance, the team hopes to free up resources on the robot for more advanced types of maneuvers.
“We are essentially finding these general principles of obstacle traversal so eventually we hope that robots can, without really needing to think much about how to traverse an obstacle like a gap or a bump, just know what to do inherently,” said Gart.
Co-authors of the journal article included graduate students Changxin Yan and Ratan Othayoth and undergraduate Zhiyi Ren, all from the Department of Mechanical Engineering.
The research was funded by a Burroughs Wellcome Fund Career Award at the Scientific Interface, a U.S. Army Research Office Young Investigator Award, and Johns Hopkins University's Whiting School of Engineering.
What do you think? Will cockroach-inspired robots support search-and-rescue? Share your thoughts below.