Could you build a robot to search for and collect samples on Mars? Team Cataglyphis from West Virginia University did. They won $100,000 last June from a prize pool of $1.5 million and are being honored in the US Senate on September 21st. The team, along with many others, have competed each June since 2012 during the NASA Sample Return Robot Challenge held at Worcester Polytechnic Institute in Worcester, Massachusetts.
The SRR, as it’s called by the teams, is a two phase competition. In Phase 1 the robot must leave the starting platform, collect a pre-cached sample, and return the sample to the starting platform. Phase 2 is more difficult because the robot must not only collect the pre-cached sample but search a park for 9 additional samples. The park is a typical urban park about 1.5 football fields large with grass, trees, and park benches as obstacles.
Since the robots are supposed to be on celestial bodies lacking magnetic fields like Mars or the Moon, they cannot use a magnetometer (compass) or GPS satellites to determine their pose, i.e. orientation and location. Add to that handicap grueling time limits of 30 minutes for Phase 1 and 120 minutes for Phase 2 and you’ve got a huge challenge on your hands.
The Mountaineers, as they were known in the robot pits, are the only team to collect two samples during the competition. Another team from Los Angeles, Team Survey, was the first to complete Phase 1 in 2013, but only managed, in 2015, to collect the pre-cached sample during Phase 2.
All the teams who have competed are waiting to see if there will be a competition in 2016 and I am among them. After the break you’ll find a couple of videos of the 2015 competition. One is about the Mountaineers but the other us from NASA 360. If you look quickly during the opening sequence of the NASA 360 video you’ll see two small black robots. One is on its side spinning its wheels; the other jammed under a rock. Those are my rovers from the 2013 SRR. I’m chasing the dream of a winning extra-planetary rover and you should too!
So dense optical flow, wheel encoder odometry, and gyroscope(s) then?
Does it say what algorithm they used for sensor fusion?
All the teams shared a lot of information but I don’t recall hearing how WVU did their sensor fusion. I was interested to see the lidar on their robot this year. I don’t recall it from ’14.
the article says they are robot team but why are all the photos and videos of a bomb?! ;)
The Mountaineers, as they were known in the robot pits
You do know why, yes?
Yes, the school mascot is the Mountaineer. I met them in ’14. I didn’t compete but since I was an old timer got access to the pits. I was not at all surprised they did well this year.
FYI: Students from Bialystok Technical University in Poland made IIRC 5 rovers like that since 2010.
so has WPI, hence why they host the competition
I’ve been surprised at the lack of university teams at the SRR. There have been students from schools but only a few university sponsored teams. UWV, WPI, and Waterloo University are the only schools who have been there, as best I recall. Both WPI and Waterloo were only demonstration teams because they were the host school or non-US. You had to be a US entity to actually win the NASA money. Waterloo did a very nice 35 page paper that was published in Jan ’14 issue of Journal of Field Robotics. It was free for awhile but best I can determine is now behind a paywall. It presented an analysis of the problem and their approach to solving it.
Oregon State also competed.