Mathematician and Associate Professor at Louisiana State University Michael Malisoff and the team he is working with are creating robots that can act somewhat autonomously, working in the place of humans in harmful places — oil spills and mosquito-infested lagoons. Instead of a human doing the dangerous work, the bots could “detect dangerous substances in dangerous places” and relay the information back to a human worker on shore.
“These robots are operating in potential hazardous situations,” Malisoff says. “It keeps the human in the loop but out of harm’s way.”
The team of students is directed by Fumin Zhang, an assistant professor at Georgia Tech. Malisoff is collaborating with Zhang for the project.
After the Deepwater Horizon BP oil spill in 2010 in the Gulf of Mexico, Malisoff, who specializes in mathematics dealing with control processes that apply to robotics, tapped the National Science Foundation and asked if self-directing robots might be useful to explore the damage. They answered in the affirmative and awarded the team a one-year grant to work on the bots and find out how the machines could be used in the future. The robots could be used to monitor the biological impact of such a spill.
“The main novelty of the research was in our use of a technique called ‘automatic control,’” Malisoff says. “The robot is able to sense where it is.”
There are four robots total; three can go underwater. ROV Beta is the name of one of the underwater robots. It can reach a depth of 100 meters and has a battery life of 10-12 hours, running on a battery bank on shore.
A feedback sensor on the robot allows it to determine its current location and the desired route to get to its destination. Unlike a remote-controlled car, this robot can figure out its current location using built-in commands and decide its next course of action using tracking control features. Underwater robots like these still require human intervention and the team doesn’t expect to make them completely autonomous, but with some fine-tuning the robots can spare people from doing dangerous work.
“Typically with marine surveys, the water is too vast, so it has to do some routing,” he explained. “Feedback control helps it put itself back on course if tides take it off course.”
The team had a test run with the bots in the Gulf of Mexico. The underwater robot was attached to a leash on the shore and would navigate into the water.
Right now, the robots don’t communicate as quickly as the team would like them to. Murky waters and the density of the ocean cause the robots to occasionally be unresponsive to feedback — like when you’re in your car and the GPS stops working when you go through a tunnel, Malisoff explained.
“Overall, we had promising results, but there’s always room for improvement,” he says. “It’s very much a work in progress.”
Three of the four robots are built by students who are overseen by the team of professors. Right now, the robots are disassembled while the team works to meliorate the problems. But students are busy, Malisoff said. Between other classes and projects, the bots are not being attended to as if it is someone’s full time job. Fortunately, the new grant runs through 2014 and provides funding for theoretical work that makes the robots perform better.
“This is one of the most exciting projects I’ve ever worked on,” Malisoff says. “I can see this has potential to help people in the Gulf Coast area in case of an oil spill or natural disaster.”
Posted by Janine E. Mooney, Editor
January 24, 2012