Re: Control electronics
Agreed: the autopilot is going to need inputs to work with. In addition to pitch, roll & yaw, it'll also really need GPS (for lat, lon & altitude) and airspeed.
Pitch, roll, yaw & GPS could be handled by a smart-phone, which could also run the autopilot/FCS s/w, but airspeed could be a bit trickier; dunno if there are already suitable airspeed sensors available from the RC world and, if so, how you would feed the input into the s/w.
Actually, if the airframe FDM is really accurately modelled I think you might be able to derive the airspeed by comparing the actual pitch, roll & yaw rates against the corresponding pitch, roll & yaw control outputs i.e control surface deflections, set in the context of air density, which is established via the GPS altitude (basically, comparing what you're doing with what's actually happening as a result of what you're doing, and comparing that with what _should_ be happening as a result of what you're doing - at some point along the curve there will be a correlation, and that's where you actually are). A real airspeed sensor would still be a better idea though, as it gives the autopilot/FCS real data to work with.