Epic. Cross state legality will probably slow progress for this but once you get support and attention I hope people will rally to your cause.
3,000 miles is quite long :) but are you intending to do it in one trip or stages?
Most of the autopilots out are now up to doing this but For 3,000 miles might be difficult. Your focus needs to be on efficiency of the airframe. If your can get that optimised then you will go further faster per hour.
Composite construction, glider-esque design is what your airframe needs to be but fitting the autopilot could be difficult. In particularly thin circumstances you could use the an open source autopilot such as APM to re-develop it into the form factor you require enabling you to use the codebase without worry and still fit it all in.
I've been thinking again. You own an Easystar already and can fly it safely so why not use it for the research. APM is the cheapest autopilot at the moment and is open source (which could be useeful later on for a thinner version) but more importantly you can catalog with during set flights.
If you use a current sensing setup (inductive or direct) and log the current during a set size test flight (2km or so) you will have data on the current setup effectiveness. Here is the good bit, you could change the speed for the flight (easily accomplished) and do many tests of this cruising flight. You will then have flight data for varying currents (speeds) for that particular motor. You can then find it's most efficient current (speed). All you do now is get a new motor and do the same. You will then have masses of data to work from to pick the best setup and once you have found a more suitable glider design you have data to try to correlate.
I would suggest having all the control surfaces you can too as it will make it easier. I reckon with sponsors you could this inside the year xD
Sounds like we have the same interest, I love data mining and using it to analyze what is truly happening. I also like the idea of focusing on efficiency/performance to ensure that I make the most of the "airtime".