Accelerometers are bound to detect such adverse effects. Correcting roll and pitch automatically shouldn't be hard. Also, I would guess that the higher the ma

Accelerometers are bound to detect such adverse effects. Correcting roll and pitch automatically shouldn't be hard. Also, I would guess that the higher the main rotor (compared to the centre of mass), the more susceptible the device. So, if you build the main rotor as close to possible as the centre of mass, you should have a more neutral design. While we're at it, I've flown a little toy with two main rotors, rotating in opposite directions, controlling power and yaw, while the tail rotor controlled pitch (the whole thing was auto-stable, so there was no need for roll control). It worked like a charm. So, if we're willing to suffer the complexity of an actual swashplate[1], we could try a design with 2 main rotors, one above, and one below, hereby getting a neutral, manoeuvrable, and (hopefully) efficient design. [1]: https://en.wikipedia.org/wiki/Swashplate_%28helicopter%29 https://en.wikipedia.org/wiki/Swashplate_%28helicopter%29