I need a denormalized body wrench setpoint (force in N, torque in N·m) in FRD, computed from post-mixer actuator commands.
Current approach:
- Read
actuator_motors(normalized[-1, 1]per rotor) - Use rotor geometry (
CA_ROTOR*_PX/PY/PZ,CA_ROTOR*_AX/AY/AZ) and effectiveness matrix to compute net body wrench - Result is still normalized
The problem:
To get SI units, I need to convert normalized actuator commands to physical thrust/torque. I see CA_ROTOR*_CT (thrust coefficient) with Thrust = CT * u^2, but the units of CT aren’t clear, and I’m not sure if it’s intended for SI conversion.
Questions:
- Has anyone converted post-allocation actuator commands to SI body wrench (N, N·m)?
- What’s the recommended way to denormalize? Should I:
- Add new parameters like
CA_ROTOR*_THRUST_MAX_N? - Use
CA_ROTOR*_CTwith documented SI units? - Use another existing mechanism?
- Add new parameters like
- Are there vehicle-specific considerations (ESC protocol, thrust curves, etc.) I should account for?
Context: This is for an EKF that needs the allocated actuator wrench setpoint (post-mixer, pre-motor dynamics) as a process input. The signal should reflect saturation/limiting from the allocator.
Any guidance or examples would be appreciated.