Questions about the crazyflie model and simulation

[xchl19]

Hello,

Thanks for your great work of developing the quadrotor simulatior.

I have several questions about the crazyflie model and simulation.

1. When I check the figure of thr drone in paper and official site of bitcraze, I noticed that the order of motors are different (in the reverse order). When checking the _physics() in BaseAviary.py the sign of z_torque is different from paper https://arxiv.org/pdf/1608.05786, I wonder if the rotation direction of motors is different from the official website?

2. I am not familar with the Euler angles in pybullet. I wonder in this simulator, what is the order of rotations axes, is it intrinsic or extrinsic rotation?

3. When I use dyn as physics, the drones fell down to underground. Are there bugs in the dynamic?

[JacopoPan]

Hello @xchl19,

1. the implementation here is mainly taken by the system identification works mentioned in the URDF files

<carlos url="https://arxiv.org/pdf/1608.05786.pdf" />
<julian url="https://www.research-collection.ethz.ch/handle/20.500.11850/214143" />
<mit url="http://groups.csail.mit.edu/robotics-center/public_papers/Landry15.pdf" />

not Bitcraze documentation directly, if you are using a specific firmware, you could indeed have to remap the inputs as it is done in BetaAviary and CFAviary

gym-pybullet-drones/gym_pybullet_drones/envs/BetaAviary.py

Line 258 in 3d7b12e

remapped_input = np.array([[

gym-pybullet-drones/gym_pybullet_drones/envs/CFAviary.py

Line 633 in 3d7b12e

def _powerDistribution(self, control_t):

2. see the "Transforms" section in PyBullet Quick Guide for the rotation order

3. running which one of the example?

[xchl19]

Thanks a lot for your reply!

About question 3, I changed DEFAULT_PHYSICS = Physics("pyb") in pid.py into DEFAULT_PHYSICS = Physics("dyn").

[JacopoPan]

Thanks @xchl19, it seems you are correct and "dyn" is not working as expected
This PR was the most recent change to its implementation but even reverting it does not seem to fix the issue, I'll need to investigate a bit deeper if something about the robot description or bullet has changed