The Diplomat
September 25, 2008
Franklin & Marshall’s weekly online newsletter
Mission to Mars
Michael Anderson
Imagine robots that can think, make decisions on their own and learn from situations.
It’s not the stuff of science fiction or movies.
It’s what one Franklin & Marshall College psychology professor and his team of students and staff are building.
Assistant Professor of Psychology Michael Anderson is helping to lead a team of students, computer scientists and engineers building the next generation of rovers to explore Mars.
“We’ll have robots up and running by the end of the year,” said Anderson.
The work at F&M is funded by a $106,000 grant from the National Science Foundation, part of a larger NSF grant of $800,000, split among F&M and three universities in Maryland.
Anderson has teamed up with researchers at University of Maryland at College Park, the University of Maryland, Baltimore County, and Bowie State University to develop the rovers and prepare the technology for future missions to Mars.
“The big difference will be the brains. The current Mars rovers are remote controlled. They are not making any decisions, but have to be indirectly controlled because of a six- to 41-minute communications delay with Mars,” Anderson explained. “Our idea is to create robots that are fully autonomous. The trouble now is when a computer encounters something unexpected. The result can be unpredictable.”
For example, Anderson explained, say a rover is traversing the planet’s surface. The going is easy, but then progress slows as the surface becomes more rocky. The current Mars rover wouldn’t know how to deal with this problem, if it even noticed, unlike the rovers Anderson envisions.
“What we want is to give it flexibility for it to think for itself,” he said. “These rovers would assess the situation and come up with a plan to overcome the problem. In this case, to find out why it was moving so slowly, and figure out a way to avoid the problem in the future. For instance, maybe the terrain has different characteristics from what was expected; the rover could notice this and steer around such terrain in the future.”
To give the rovers the ability to adapt to their surroundings, Anderson said, the team must create a “meta-cognitive loop.” That means the ability to see a problem, evaluate it, make a decision, act on that decision and learn from the experience. Sounds easy. Easy for a human or an animal, he said, but not so easy for a machine.
“That will take a lot of programming, and in a special reflective, expectation-sensitive style that the team has developed over the past several years,” he said.
Once built, the rovers will work under the direction of their human creators, receiving tasks and orders from them. The controllers would communicate with the rovers and help them evaluate situations so the rovers don’t learn a lesson incorrectly.
Anderson soon will begin working with F&M students to build rovers. He hopes to have rovers driving around Barshinger Life Sciences and Philosophy building by the end of the school year.
To do this, Anderson will enlist the help of a team of students. So far, he has picked Physics major Boyko Borislavov Perfanov ’09. He’ll be tasked with evaluating various commercial robot platforms, then helping Anderson build and customize the rovers. Other students are likely to be brought into the project soon. Also assisting is F&M Electronics Engineer Steve Spadafore.
If the project is successful, Anderson and his colleagues will have developed the brains NASA could place in rovers on their way to Mars sometime in the next decade.
