MSU Robotics Club wins Penn State Mini Grand Challenge
They left with trepidation and doubt, but they returned as champions.
The Minot State University Robotics Club recently won first place in the Penn State Abington Mini Grand Challenge outdoor robot contest, held April 16 in Abington, Pa. MSU associate professor Larry Atwood and assistant professor Scott Kast, a team of four students and robot contestants Einstein and Navigator drove two days and over 1,700 miles to participate in a competition that entailed a mere half-mile journey across a suburban campus. Only Einstein and Navigator weren’t worried.
The journey didn’t start there though; it actually began in January.
"We began preparing for this competition around Jan. 15," Atwood said, "but we really could’ve used another month. That would’ve allowed us time to practice outside and on grass."
Mallard geese migrate thousands of miles to Panama, Cuba and the Bahamas every year. So what makes navigating a half mile across campus so difficult? It is not that easy.
First offered in 2005, the Penn State Abington MGC outdoor robot contest challenges mobile, ground robots to autonomously navigate paths throughout the campus of Penn State Abington College while avoiding obstacles and tackling off-road detours, without communication or remote controls. The robots must also transport one gallon of water to display payload capabilities, and they need to "entertain" spectators. Only one robot to date has successfully completed the entire course.
"The Penn State path was asphalt and there was no snow on the ground, unlike here, which made it hard to prepare for," Kast said. The path also had green grass, wood chips, rocks, cones, a lake to navigate around, a bridge to cross and people periodically jumping out in front of the robots testing their ability to recognize unanticipated obstacles.
"They also used shorter cones," Atwood added, "than what we used in practice."
The objective of the contest is to promote advances in engineering design, computer technologies, artificial intelligence and robotics education, according to the contest website. This year a total of seven teams competed; six schools and one team of engineers.
"I was feeling confident until two days before our park trial," Kast said. "Up until then we had practiced inside on carpet and cardboard trails and thought our (software) code for the path was doing great."
After the first outdoor trial, Kast realized they had a lot of adjusting to do. In addition to coded computer software, the team utilized cameras, a compass, sonar, and GPS to determine the final destination. They also used online pictures to analyze the campus path.
"This is very much a team effort," Atwood said.
The Advanced Robotics class represents a semester of student work that includes building robots and writing computer code to interpret images and ground surface, detect objects, avoid obstacles, travel the intended path and stop at the end. Atwood and Kast have taught the class for five years.
"Larry suggested we go early to practice on the course, which was a good idea," Kast said. "The second-place robot made it down a hill, but when someone jumped in front of it, it turned to face a building and then never moved again. We were able to change our code to avoid that."
When MSU took its turn on the course, Navigator had technical difficulties with its sonar and shut off, which eliminated it from the competition. That left Einstein to win the day, securing first place, a trophy and the $500 prize.
Even though MSU won, Atwood was disappointed because the team wanted to complete the whole course during the competition. Ironically, Navigator unofficially traversed the entire course over its next two runs.
Kast said it would be fun to attend again next year. Atwood isn’t so sure. Looking back, though, Atwood’s perspective has softened.
"We had the most sophisticated robots there," Atwood said.