The two hardware aspects of the competition that pose a challenge are the means of collecting a sample and the 'sterile' transport of them to the starting platform.
For the most part, there are commercial off-the-shelf (COTS) platforms that are available or could possibly be up-scaled to provide the basic locomotion needed for the contest. this is based on my initial thoughts from the first description of the competition roving area. It sounded like a park or maybe a golf course setting. The subsequent satellite and other photos, along with the topographic map upheld that suspicion. While New England is known for its rocks - see Robert Frost's "Mending Wall" - there was no indication that large rocks would be encountered.
The Robot Shop contains a number of suggestive platforms, although they would need up-sizing for the competition:
CoroWare CoroBot in various permutations, some including a possible arm.
CoroWare Explorer in various permutations
GEARS SMP Mobile Platform is an example but would need up-sizing
Dr. Robot Jaguar 4x4 Mobile Platform is a bit small and overkill on the power but is illustrative of other 4-wheel solid platform models
Dr. Robot Jaguar Lite Tracked Mobile Platform is, again, small but shows a tracked version
For lack of a better term, I will use 'pickers' to describe the mechanism that takes a sample off the ground and gets it onto the robot.
Collecting the samples is one challenge where COTS may not provide a solution, primarily in picking up the 'hard' advanced samples. These three samples are non-ferrous metallic objects each with a unique engraving. (For some reason I have the idea these a flat disks. If so, this poses a real challenge for picking them up. Can anyone elaborate?) While the first thought is to use a pincer hand, that may not be the best approach. The sphere or hockey puck may easily slip if not aligned perfectly.
Looking at the photos from the June competition there are a number of different pickers. One was the usual pincer hand that is available from a number of sources.
The SpacePride team has an interesting set of grippers. They are two horizontal, circular fingers that close in from the sizes to catch the sample. The sample is lifted straight up and the 'hand' moves on the X-Y axis to drop the sample into a bin.
The Windsor team's robot had something like a plow that dropped down and a broom pulled the sample onto the broom. The plow lifted and the sample tumbled into a bin.
I can't tell too much about the other team's pickers and I may be wrong in detail about the ones I did describe.
My thoughts last year were along the lines of the Windsor team or along the lines of a claw.