ci(pre-commit): autofix

Author: pre-commit-ci[bot]

docs/tutorials/mapping_based_calibrator.md

@@ -147,7 +147,6 @@ The image below displays the vehicle within the pointcloud, allowing for a compa
 ## FAQ
 
 - Why does the calibration fail?
-
   - In most cases, the failure is due to mapping issues. The possible error messages are listed below and should be displayed in the console. For these cases, restart the experiment and drive more stably and slowly.
     - Mapping failed. Angle between keyframes is too high.
     - Mapping failed. Interpolation error is too high.

docs/tutorials/marker_radar_lidar_calibrator.md

@@ -231,19 +231,16 @@ To evaluate the calibration result, the user can check that the calibrated radar
 ## FAQ
 
 - Why does the reflector detection not appear on `RViz`?
-
   - Make sure the center of the reflector faces toward the radar sensor, and the height of the reflector matches the radar's.
   - Make sure the height of the radar reflector is not larger than the `reflector_max_height` parameter.
   - Make sure the radar reflector is not in the background voxel (visualize the topic mentioned before).
 
 - Why does the calibration error seem high?
-
   - Make sure that there are no outliers in the calibration pairs list.
   - Make sure that the initial calibration is good enough to match the lidar detection and radar detection correctly.
   - Radars like the ARS408 (the one we use in this tutorial) have a resolution of 0.2m. Given that, there is a theoretical limit to how low the calibration error can be. If the radar resolution is low, it will strongly limit the lower bound of the calibration error.
 
 - When can I stop the calibration process?
-
   - It is recommended to stop the calibration when the curve in the cross-validation error has converged.
   - With more matched pairs without outliers, the calibration result should improve if the number of pairs increases.
 

docs/tutorials/tag_based_pnp_calibrator.md

@@ -133,25 +133,21 @@ In the following images we can see how the projection looks using the initial ca
 ## FAQ
 
 - Why does the tool fail to add calibration pairs?
-
   - One possible reason is that the detections are too close to the previously collected data. In this case, the new detections are deemed redundant, and thus not accepted.
   - The timestamps of the lidar and camera are not synchronized. This can be checked with `ros2 topic echo [topic_name] --field header.stamp`. Setting the parameter `use_receive_time` to `True` might help to solve the issue, but is not recommended as a long-term solution.
   - The detections are not stable enough. This can happen due to the following reasons:
     - The detection rate is not stable: this can happen when the lidar points inside the tag frames are not sufficient for the algorithm to detect them reliably, which usually happens when the tag is either far away from the sensor or outside its high-density zone (when applicable). In these cases, please move the tag to a position where enough points (scan lines) hit the tag. If the user forcefully calibrates under undesirable conditions, the results can get compromised.
     - The tag is physically unstable due to wind, mounting issues, or other external factors. Even if the detector functions correctly, these conditions prevent the detection from converging. If this is the case, please eliminate all unwelcome external factors before attempting calibration. Forcefully calibrating under these conditions can compromise the results.
 
 - Why does the UI not launch?
-
   - Check with `ros2 node list` if the relevant nodes have been launched. It is possible that the provided parameters do not match any of the valid arguments among other standard ROS issues.
   - If the UI crashes (check the console for details), it is probably due to a bad PySide installation, invalid intrinsic parameters, invalid extrinsic parameters, etc.
   - The timestamps of the lidar and camera are not synchronized.
 
 - Why does the reprojection error increase when more data is collected?
-
   - When there are few samples, the model will fit the available data the best it can, even in the presence of noise (over-fitting) or wrong detections. When more data is collected, the error may increase to a certain extent, but that corresponds to the model attempting to fit all the data, this time unable to fit the noise, resulting in a higher error. However, it should reach a more-or-less table peak with about 10-15 pairs (depending on the data collection pattern/sampling).
 
 - Why does the reprojection error seem high?
-
   - The intrinsics may not be accurate, thus limiting the performance of the method.
   - The boards are not appropriate (are bent).
   - The boards moved too much while calibrating.