When you think “quadrupedal robots”, Boston Dynamics’ SpotMini might be the first thing that comes to mind. But while Spot has been shaking it to Bruno Mars, a competing robot dog has started training to do the dirty jobs no human wants to do.
Say hello to ANYmal, a four-legged machine capable of autonomous exploration that will one day venture through all 572 miles of Zurich, Switzerland’s famously golden — but no less gross — sewer system. Its creators, the Swiss Federal Institute of Technology (ETH Zurich) and robotics company ANYbotics began the tests on December 19. The robot’s onboard tech allows it to take pictures and measure the geometry and temperature of sewage tunnels for routine inspections. It’s necessary work that humans have had long had to do.
ANYmal’s head is retrofitted with an array of cameras and Lidar laser sensors, which are used in autonomous cars. The mechanic pup is is capable of autonomous travel in laboratory conditions, but ANYbotics co-found Peter Fankhauser and his team have decided control it with a joystick initially.
“It’s a precautionary measure,” Fankhauser says. “Just because something works in the lab doesn’t always mean it will in the real world.”
Fankhauser and his colleagues are participating in three-year research project titled THING (sub-Terranean Haptic InvestiGator). The end goal is to make robots, like ANYmal, to understand and navigate through their surroundings. ANYmal’s creators want to give it touch-sensitive feet so it can detect on anything going on underneath it.
Robots like this can alleviate a hefty burden on workers that are tasked with ensuring crucial infrastructure, like sewage systems, are always working. In 2017, a stationary engineer for the New York City Department of Environmental Protection worked an average of 78 hours a week. That’s almost double the average 40-hour week you’d work if you have a nine-to-five job.
While ANYmal wouldn’t replace engineers, it could make sure everything is running smoothly and call on human assistance when something goes wrong. This way humans can spend less time monitoring and more time refining their specialized skills.
If THING is a success, urban infrastructure workers could have a pack of robot dogs they let loose underground.