My initial though is that there is a lot of system complexity. Lots to develop, lots to go wrong. But...
On e-paper i think variable speed prop makes sense, since you can keep the blade closer to optimum AOA. The ideal would be active propeller twist so that each prop element could run at optimum AOA. I thought the Carter Copter variable pitch prop an interesting approach. The PRPM would be altered depending on required thrust, and likely mapped to reduce PRPM with speed.
In practice i can see many problems with V-PRPM. For a start if the prop (or twin props) is/are also being used for yaw control the system gets very complex very quickly. In this case altering PRPM with airspeed only (if at all) is likely the best solution.
I actually think this is an area where maybe the hybrid helicopter has a major advantage - as long as no individual motor failure puts the machine at instantaneous risk (rotor clash, pilot overload etc). An important rule with hybrid systems is don't compromise your prime mover power/weight (ie force it to deliver high power at low RPM).
Regarding the proposed system, i can see it working only in situations of perfectly balanced torque (ie first time you bank it stops working
). For automotive applications differentials are a necessary evil (but interesting to develop
). They attempt to balance torque for potentially unknown velocity differences. This is the problem though: when the torque no longer balances the diff no longer works. In your case the servo motor would end up having to introduce/remove quite a lot of energy. I've done calcs for similar active differential concepts, and found that a 20kW motor/gen would be required for my own FWD 1 tonne car!
A better approach would be to drive rotor and prop through seperate gearboxes, and vary the ratios to suit. CVT units are out there now, with more in development...
Mart
[Edit: not yet, but i'll bet i think of somthing after a coffee..]