For example, ultrasound is already used to wirelessly charge pacemakers:
http://www.economist.com/blogs/babbage/2011/11/wireless-pace...
And here is a very pertinent patent in this space (2004) that also contains a nice review of prior art:
http://www.google.com/patents?id=508SAAAAEBAJ&zoom=4&...
The central difficulty in this idea, as pointed out in the above patent, is automatically aiming a sufficiently focused beam at the receiver. Otherwise you get killed by 1/rˆ2 power decay. And this is a non-trivial task, which will force them to either go the royalties way or dodge the existing patents.
I don't doubt the team can move forward, but I wonder if the undergrads (and their investors) have underestimated the difficulty and history of the problem.
That being said, it's refreshing to see folks inspired by real-world problems that don't involve "social".
Wishing them best of luck!