New Crowdin translations - ko (#26487)

Co-authored-by: Crowdin Bot <support+bot@crowdin.com>
This commit is contained in:
PX4 Build Bot
2026-02-18 16:44:57 +11:00
committed by GitHub
co-authored by Crowdin Bot
parent 602add3ec1
commit 18c176beef
314 changed files with 11938 additions and 1357 deletions
@@ -1,13 +1,68 @@
---
pageClass: is-wide-page
---
# AutotuneAttitudeControlStatus (UORB message)
Autotune attitude control status
Autotune attitude control status.
This message is published by the fw_autotune_attitude_control and mc_autotune_attitude_control modules when the user engages autotune,
and is subscribed to by the respective attitude controllers to command rate setpoints.
The rate_sp field is consumed by the controllers, while the remaining fields (model coefficients, gains, filters, and autotune state) are used for logging and debugging.
[source file](https://github.com/PX4/PX4-Autopilot/blob/main/msg/AutotuneAttitudeControlStatus.msg)
**TOPICS:** autotune_attitudecontrol_status
## Fields
| 명칭 | 형식 | Unit [Frame] | Range/Enum | 설명 |
| ------------------------------ | ------------ | ---------------------------------------------------------------- | --------------- | -------------------------------------------------------------------------------------------------------------------- |
| timestamp | `uint64` | us | | Time since system start |
| coeff | `float32[5]` | | | Coefficients of the identified discrete-time model |
| coeff_var | `float32[5]` | | | Coefficients' variance of the identified discrete-time model |
| fitness | `float32` | | | Fitness of the parameter estimate |
| innov | `float32` | rad/s | | Innovation (residual error between model and measured output) |
| dt_model | `float32` | s | | Model sample time used for identification |
| kc | `float32` | | | Proportional rate-loop gain (ideal form) |
| ki | `float32` | | | Integral rate-loop gain (ideal form) |
| kd | `float32` | | | Derivative rate-loop gain (ideal form) |
| kff | `float32` | | | Feedforward rate-loop gain |
| att_p | `float32` | | | Proportional attitude gain |
| rate_sp | `float32[3]` | rad/s | | Rate setpoint commanded to the attitude controller. |
| u_filt | `float32` | | | Filtered input signal (normalized torque setpoint) used in system identification. |
| y_filt | `float32` | rad/s | | Filtered output signal (angular velocity) used in system identification. |
| state | `uint8` | | [STATE](#STATE) | Current state of the autotune procedure. |
## Enums
### STATE {#STATE}
| 명칭 | 형식 | Value | 설명 |
| ---------------------------------------------------------------------------------------------------------------------------------------------- | ------- | ----- | ------------------------------------------------------------------------------------------- |
| <a href="#STATE_IDLE"></a> STATE_IDLE | `uint8` | 0 | Idle (not running) |
| <a href="#STATE_INIT"></a> STATE_INIT | `uint8` | 1 | Initialize filters and setup |
| <a href="#STATE_ROLL_AMPLITUDE_DETECTION"></a> STATE_ROLL_AMPLITUDE_DETECTION | `uint8` | 2 | FW only: determine required excitation amplitude (roll) |
| <a href="#STATE_ROLL"></a> STATE_ROLL | `uint8` | 3 | Roll-axis excitation and model identification |
| <a href="#STATE_ROLL_PAUSE"></a> STATE_ROLL_PAUSE | `uint8` | 4 | Pause to return to level flight |
| <a href="#STATE_PITCH_AMPLITUDE_DETECTION"></a> STATE_PITCH_AMPLITUDE_DETECTION | `uint8` | 5 | FW only: determine required excitation amplitude (pitch) |
| <a href="#STATE_PITCH"></a> STATE_PITCH | `uint8` | 6 | Pitch-axis excitation and model identification |
| <a href="#STATE_PITCH_PAUSE"></a> STATE_PITCH_PAUSE | `uint8` | 7 | Pause to return to level flight |
| <a href="#STATE_YAW_AMPLITUDE_DETECTION"></a> STATE_YAW_AMPLITUDE_DETECTION | `uint8` | 8 | FW only: determine required excitation amplitude (yaw) |
| <a href="#STATE_YAW"></a> STATE_YAW | `uint8` | 9 | Yaw-axis excitation and model identification |
| <a href="#STATE_YAW_PAUSE"></a> STATE_YAW_PAUSE | `uint8` | 10 | Pause to return to level flight |
| <a href="#STATE_VERIFICATION"></a> STATE_VERIFICATION | `uint8` | 11 | Verify model and candidate gains |
| <a href="#STATE_APPLY"></a> STATE_APPLY | `uint8` | 12 | Apply gains |
| <a href="#STATE_TEST"></a> STATE_TEST | `uint8` | 13 | Test gains in closed-loop |
| <a href="#STATE_COMPLETE"></a> STATE_COMPLETE | `uint8` | 14 | Tuning completed successfully |
| <a href="#STATE_FAIL"></a> STATE_FAIL | `uint8` | 15 | Tuning failed (model invalid or controller unstable) |
| <a href="#STATE_WAIT_FOR_DISARM"></a> STATE_WAIT_FOR_DISARM | `uint8` | 16 | Waiting for disarm before finalizing |
## Source Message
[Source file (GitHub)](https://github.com/PX4/PX4-Autopilot/blob/main/msg/AutotuneAttitudeControlStatus.msg)
:::details
Click here to see original file
```c
# Autotune attitude control status
@@ -55,5 +110,6 @@ uint8 STATE_TEST = 13 # Test gains in closed-loop
uint8 STATE_COMPLETE = 14 # Tuning completed successfully
uint8 STATE_FAIL = 15 # Tuning failed (model invalid or controller unstable)
uint8 STATE_WAIT_FOR_DISARM = 16 # Waiting for disarm before finalizing
```
:::