That concept pic looks awesome.
...and you could probably achieve it with fibreglass sheets and resin. Pretty cheap stuff, usually used for making/fixing model helicopter bodies. Could make for an awesome monocoque airframe, leaving plenty of room inside for rocket engines. Will you be just buying off-the-shelf Estes stuff, or making your own engines? Also are you considering solid, liquid or hybrid rocket engines? From what I recall, the Top Gear shuttle attempt used latex and nitrous oxide. Could be worth contacting Rocket Men Ltd to see if they'd like to join forces or find out how much it'd cost to hire their expertise.
As you're going above GPS heights, I'm going to guess you'll need to work without it, at least at the start of the launch. Solid state barometers can be fettled to work as altimeters, and you already know about solid state gyroes and accelerometers. Problem is, gyroes will tend to drift so you will need to find a way of detecting absolute orientation every now and then (as opposed to the gyroes just telling you how much you rotated since the last measurement). You could go for optical horizon detection (probably a good option as you'll be way above cloud level and presumably launching in the day, even if it would be computationally expensive to analyze a couple of webcam feeds). Another more iffy method is magnetic detection. Apparently the difference in the Earth's magnetic field is measurable over a pretty short altitude difference, so sensors in the wingtips could detect whether you're flying level or not (at the expense of not knowing if you're upside down or not).
Another (better) method of horizon detection is putting IR sensors in the wingtips and looking for a warmer earth versus a cooler sky. This is used in auto-land mechanisms that are already available, however they have their own drawbacks. You'd need extra sensors above/below to detect inversion, and the behaviour of an autoland system might get a little odd if it approaches a tree line or hilly terrain. I've seen some reports saying the aircraft will automatically bank away from the tree line, but a sharp bank at low altitude will likely end up with a nosedive into the floor.
Have you thought about possibly taking manual control of the aircraft once it gets low enough, presuming you can find the thing in the sky? Can we have a re-attempt at trying to track the craft via telescope? Also, please put a forward-facing camera on the aircraft this time. PARIS was awesome in very many ways, but it was a shame to not be able to see the ride down from the Playmonaut's perspective.
Also, perhaps it's possible to have the engines as external modules that get jettisoned after exhaustion? Extra brownie points for having them come down via parachute, shuttle-SRB-style. Extra extra brownie points for sticking cameras on them too.
For location, maybe you could use a lightweight, hacked smartphone with GPS in addition to the radio beacons? It's likely the batteries would freeze at very high altitude, but so long as it wakes back up properly once things warm up a little (and presuming it comes down in an area with some signal) you'd have a very precise way of recovering your spaceplane (and perhaps SRBs).
Maybe put a radio beacon on the floor somewhere and see if the plane can home in on it?
My word, this post is getting long. Hopefully not too long. Dammit Lester, there's just so much stuff to think about here. I'm envious you're getting to do it at all though!
Biting the hand that feeds IT
