Hello. I’m a complete newbie here and I am brimming w/ questions.
First a few words about myself. I got into the RC aircraft (fixed wing & heli) many years ago. I got out of the hobby but now I have more free time than before. I recently bought 2 small electric helis, a BLADE MCS which I fly around the house and a small BLADE quadrocopter which I fly outside when there isn’t much wind. When I first saw a quadrocopter years ago (a Draganflyer) I was really impressed by the design and I think I figured out the basic physics of how it flies. If I’m not mistaken, the basic idea is this:
The fore & aft rotors turn CW & the right & left rotors turn CCW. In equilibrium (no rotation, no change of altitude, no movement right or left or forward or back) the thrust of the 4 rotors just equalizes g & the net torque is zero. An accelerometer (A-RL) set up along the RL axis and another (A-FA) along the forward-aft axis. If the copter tips a bit to the right, the A-RL senses this and tells the controller, which sends a bit more juice to the right rotor & a bit less to the left. And similarly for tipping in the other directions. If the pilot move the right stick a bit to the right, the receiver picks this up & sends the info the controller, which decreases current in the right rotor & increases current in the left rotor and the copter pitches to the right until the angle w/ the horizontal, perceived by the (A-RL) corresponds to the deflection of the transmitter stick to the right. Now the plane of thrust is pitched to the right, so the thrust vector has a component pointing to the right and the copter accelerates in this direction. And similarly for the other directions. To go up give all 4 rotors more current, to go down less. To rotate the copter CW, give the right and left more current & the fore and aft rotors correspondingly less. Do the opposite to rotate CCW.
Question1) Is the above paragraph more or less correct?
Question2) If I were to buy the Ardupilot Mega-electronics set from UDrones, would I get the ability to examine and modify the source code on my PC?
P. S. I'm not sure how to use this forum. was "add Discussion" the right way to go?
Replies
George,
For 1), that is close to the right idea. The x and y accelerometers on a quad don't actually sense much on a quad because as it tilts, it also accelerates in a way that negates the acceleration of gravity to the sensor. Another way to think about this is that the accelerometers measure external forces on the quad, and the thrust vector is always directed perpenicular to the x and y accels, but they would pick up horizontal acceleration caused by a gust of wind, for example. To overcome this, the APM also uses angular rate gyros, which are integrated and combined with accelerometer data to form a total estimate of the quadrotor attitude, and this is used to drive the motor to stabilize the quad.
With respect to 2), the Attitude.pde and motors_quad.pde code files are where you can find the core of the stabilization and motor mixing control code.
1) Yes!
2) Yes! It's 100% open source. There's no need to do any programming at all, but if you want to look at the source code (or modify it), you can. It's here.