November 26, 2012
Autonomous robots reducing human risk
Autonomous robots that can perceive and react to the ever-changing scenarios have the potential to tackle situations too dangerous for humans.
There are some things that humans simply shouldn’t have to do, says Steven Waslander, an assistant professor in the Department of Mechanical and Mechatronics Engineering at University of Waterloo, and founder of the Waterloo Autonomous Vehicles lab (WAVElab.)
“Robots can do the things that are dangerous, boring or messy,” he says — things like cleaning up a nuclear plant after it’s been pounded by a tsunami, or doing aerial inspections of pipelines in remote locations. WAVElab controllers are part of Aeryon Lab’s Scout drone, which is used for aerial inspections by police, military and marine organizations, and in global hotspots including Libya. Waslander also worked with a group that used robotics to get a better understanding of the aftermath of the Fukushima nuclear disaster.
“We’re basically trying to get robots to be totally autonomous,” he says. “We’re trying to eliminate all situations where humans are in danger.”
The challenge is developing an autonomous device that can perceive and react to the ever-changing scenarios that exist in the real world. The work continues. Building a robot that can interact with fixed objects is, by comparison, child’s play. “It’s harder when they things around you aren’t controlled,” says Waslander.
Think about remote-controlled drones that dive from Middle Eastern skies on unsuspecting terrorists. Take it a step further, to a robot with the perception of a human and the ability to communicate those to other robots and to its human handlers.
“There is a huge military application,” Waslander says, “but we focus on civilian applications like inspection tasks, search and rescue, and surveillance and tracking.”
On his website, Waslander describes his work as the development of perception, planning and co-ordinating algorithms for autonomous vehicles. His primary focus, since his days of graduate study at Stanford University, has been on aerial vehicles, such as quadrotor helicopters and Micro Air Vehicles, but his student teams do work on ground vehicles as well.
Waslander’s interest in autonomous vehicles took root after during his undergraduate years at Queen’s University where he was in an applied mathematics program.
“I stumbled into this,” he says. “When I was at Queen’s I took a controls course, and I realized you could use mathematics to model and control these systems reliably.”