The birth of an idea
This is a description of my idea for a new type of aircraft. All aircraft fly in 3D space, but must fly within an envelope determined by their construction. Airplanes (generally) can only fly forward into a conical volume. Helicopters can translate in a spherical volume but can only rotate (mostly) on their vertical axis. The craft I will describe can translate and rotate about all axes simultaneously and independently.
This idea is an offshoot from flying tricopters and quadcopters. These radio controlled craft are becoming possible and popular due to the development of low cost three axis accelerometers, magnetometers, and rate gyros. With these sensors, low cost and powerful processors, and efficient programs to calculate heading and orientation, stable flight can be achieved in these inherently unstable aircraft configurations.
But for some reason I wanted something more, something unique. My friends at work were flying RC helicopters and I eventually arrived at the idea of setting pairs of tail rotors on each of the three principal planes, symmetrically about the a central point. All rotors would be driven at the same speed and only the blade pitch would change, either positively or negatively. For each plane, changing pitch together would cause translation, and differentially would cause rotation. This physical configuration, along with a sufficient power to weight ratio, and with fast enough control algorithms should be able to achieve 6 DOF flight.
With the initial idea in mind, how can I achieve this? There are a number of issues to address. The aircraft as described above seems to be the easy part. While laying out the concept in 3D CAD I saw that the craft could be constructed as a disk with six equally spaced arms with rotors at the arm ends tilted alternatingly at 55 degrees. This creates three symmetrically distributed pairs of rotors on three principal planes. Copying the shaft drive configuration and pitch control mechanisms from RC helicopter tail rotors, along with lithium batteries, motor, and controller completes the physical craft.
The control theory seems straight forward by building on the existing hobbyist software available. One author in particular has written clear and articulate articles that make the endeavor seem possible. Comparing earth referenced, craft referenced, and user input intentions and mapping the errors using direction Cosine matrices to adjusting pitch angles accordingly seems possible. I would like to contact the author to try and enlist his help.
The third part of this concept is the user input/control of the craft. The existing radio control input configurations do not seem to address the needs of 6DOF flight. I have a 3D input device for CAD work that seems a very good match for this need. The 3D Connections Spaceball input device allows 6DOF input and matches the freedom of the aircraft. I envision the user inputs are the desired acceleration for translation and velocity for rotation of the craft. With the controller at neutral the craft would hold position. But for this to work the controller also needs to be earth referenced. This is easily accomplished by incorporating an AHRS similar to the one in the aircraft into the radio controller. This allows the user to move about and know that moving the controller to the North will make the aircraft move northward. There is never any control reversal and control should be very intuitive and easy to learn. There is no throttle needed either, but for energy efficiency and longer flights, a throttle should be incorporated so if not moving rapidly the rotors can slow down some. I think of the throttle as setting how rapidly the craft can react to user inputs. If violent motion is desired the throttle is advanced else a lower setting is used.
AHRS Attitude and Heading Reference System
An AHRS, as its name implies, gives you your heading and attitude relative to the geographically local North and horizontal. More specifically it tells you the angular difference between your crafts horizontal heading and North and your roll and pitch from the local horizontal. This is very useful information, but it does NOT tell you if you are moving.
Inertial navigation uses acceleration information to continually update your position and speed in three dimensions. Simple physics allows you to solve for location and velocity if you have continuous acceleration information and know your initial location and velocity. The key is accuracy and continuity. Errors in acceleration compound rapidly so it is difficult to maintain accuracy over time and distance.
Global Position Satellite
GPS gives us relatively slow but accurate updates on location in 3D space. The system is biased towards horizontal accuracy so the vertical accuracy is lower.
To use velocity inputs from the user, the craft needs to know both its velocity, both in orientation and in translation. An AHRS only gives orientation and the GPS system gives too slow of updates for stable flight (I believe). The logical solution is to use short term acceleration data to integrate location and velocity and keep recalibrating against the GPS data. This makes high frequency location and velocity data available for aircraft stability while maintaining longer term positional accuracy. To deal with the lower vertical accuracy, a barometric sensor can be used for better short term altitude positioning. Writing this software is currently beyond me but with research and perseverance I believe it is possible.
I believe that if the rotors are all running at the same speed and each pair spin in opposite directions, gyroscopic effects will be nulled. If true, rotation on one axis will not impart a rotation on an orthogonal axis. This will make control of the craft, and the software much simpler. There will be forces imparted on the arms holding the rotors so they need to be strong and rigid.
I have seen people building variable pitch quadcopters online. None of these are 6DOF craft but they do expand the capabilities of multicopter aircraft. Some use the main rotor system from helicpors so the oneway clutch will allow the rotors to continue to windmill if the motor stops. I see no benefit for these types of craft as control will be lost so windmilling will not help. For this craft I think a single motor gives the best efficiency but I will forgo the weight of the oneway clutch. Driving all six arms from a single central gear does seem like a good approach. I would like to have some speed reduction at the rotors so the drive shafts are spinning faster to reduce their strength requirements. Ideal is a central motor spinning a bevel gear, high speed shafts, and slower rotors. I will need to model and refine this approach.
A shroud protecting the rotors would be nice. My initial idea was a sphere, but a torus would also work for this configuration. An open and rigid wire mesh would be ideal. It should also be part of the frame because weight is important as for all aircraft.
The second generation ArduPilot hardware should be able to handle the computations required and has the needed sensors. My 9 axis sensors and Arduino board can supply the AHRS information for the radio transmitter. I hope the 3D Connection device can integrate to the Arduino. I have never seen drivers for Arduino, but I never looked either.
OK, I'll stop rambling. That's my idea. Comments?