ParameterEstimation

This parameter estimation node estimates a default parameters from inputs for steer offset,wheel base and gear ratio.

I/O

input

The following topics are used to estimate the parameters.

• /sensing/imu/imu_data: used as vehicle angular velocity
• /vehicle/status/twist: used as vehicle velocity
• /vehicle/status/steering: used as vehicle steering angle (Only used in Steer Offset Estimator & Wheel Base Estimator)
• /calibration/vehicle/handle_status: used as vehicle handle angle (Only used in Gear Estimator)
• /vehicle/engage: used to check the driving operation status

output

The following topics are the output

• /vehicle/status/gear_ratio
• /vehicle/status/steering_offset
• /vehicle/status/wheel_base

For users, the EstimationResult.msg output contains the following items:

• result: Estimated result.
• result_mean: Average value of result.
• result_stddev: Standard deviation value of the estimated parameter
• absolute_error: Absolute error of the estimated parameters
• mean_absolute_error: Mean absolute error of the estimated parameters
• stddev_absolute_error: Standard Deviation of absolute error of the estimated parameters

These values can be confirmed in plot_juggler

• To display in the plot juggler, you need to drag the result into the table, currently plot juggler cannot auto display the value

How to Run Parameter Estimator

Note: You need to build the Autoware beforehand.

The following command will start the parameter estimation node.

ros2 launch parameter_estimator parameter_estimator.launch.xml vehicle_model:=lexus

• The following arguments are using to select the estimator, default is true
  • select_gear_ratio_estimator
  • select_steer_offset_estimator
  • select_wheel_base_estimator
• To deselect the specific estimator, you need to set estimator to false
• Example, the following command will only launch gear ratio estimator

ros2 launch parameter_estimator parameter_estimator.launch.xml vehicle_model:=lexus select_steer_offset_estimator:=false select_wheel_base_estimator:=false

If you want to launch with Rviz, use the following launch file. Currently unavailable

# Launch parameter Estimator with the Autoware
$ ros2 launch parameter_estimator parameter_estimator_with_simulation.launch.xml map_path:=.../kashiwanoha2/ vehicle_model:=jpntaxi sensor_model:=aip_xx1 rviz:=true

How to check the estimated parameters

The necessary information is plotted in the plot_juggler, which displays the following information from top to bottom.

• First row: Estimation results confirmation
• Second row: error for estimation

You need to adjust the value of (valid_min_) or (valid_max_). according to the standard deviation to determine the validity of the data.

Estimation results confirmation

Check the estimation results.

• result : parameter estimated at each step
• result_mean : Average of the parameters
• result_stddev : Standard deviation of the estimated parameters

It is preferable to use the _mean for the calibration results.

The parameters estimation starts when enough data is stored. The output value is zero until it is ready.

Error for parameters estimation

Check the statistics of the errors in the input/output data after the parameter estimation.

• absolute_error : absolute error of measured and estimated of the parameters estimation
• mean_absolute_error : Mean of the absolute error calculated in each estimation step
• stddev_absolute_error : Standard deviation of the absolute error calculated in each estimation step

If these values are large, the model needs to be reconsidered.

Data preprocessing

Examine the results of processing the input data

Data that do not satisfy the following conditions are considered invalid and will not be used for estimation.

• Data variation (evaluated by standard deviation) is smaller than the threshold
• The car is not in automatic operation mode (judged from Autoware’s Engage, Vehicle’s Engage, etc.)

Changelog for package parameter_estimator

0.2.0 (2025-03-24)

0.1.0 (2025-01-28)

• unify version to 0.0.0
• feat(parameter_estimator)!: replace tier4_debug_msgs with tier4_internal_debug_msgs (#198)
  • feat(parameter_estimator)!: replace tier4_debug_msgs with tier4_internal_debug_msgs

  * style(pre-commit): autofix ---------Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]\@users.noreply.github.com>

• refactor(global_parameter_loader): prefix package and namespace with autoware (#159)
• fix: some fix on calibration_adapter and parameter_estimator (#127)
  • some fix on calibration_adapter and parameter_estimator
  • minor fix

  * style(pre-commit): autofix ---------Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]\@users.noreply.github.com> Co-authored-by: Yuxuan Liu <<619684051@qq.com>>

• Changed from map.launch.py to map.launch.xml (#71)
• feat: add vehicle, control and localization calibration tools (#22)
  • add vehicle control localization tools
  • style(pre-commit): autofix
  • fix pre commit
  • Fix to pass cspell about localization
  • ignore words from spellcheck
  • ROS2 -> ROS 2
  • feat: convert autoware_auto_msg to autoware_msg (#42)
  • merge
  • delete autoware_auto

  * style(pre-commit): autofix ---------Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]\@users.noreply.github.com>

  • add prefix to vehicle_info_utils
  • style(pre-commit): autofix
  • fix: add s to VehicleInfoUtil

  * fix

  Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]\@users.noreply.github.com> Co-authored-by: Shintaro Sakoda <<shintaro.sakoda@tier4.jp>> Co-authored-by: SakodaShintaro <<rgbygscrsedppbwg@gmail.com>> Co-authored-by: tomoya.kimura <<tomoya.kimura@tier4.jp>>

• Contributors: Esteve Fernandez, KeiNakazato, Ryohsuke Mitsudome, SakodaShintaro, Yutaka Kondo, yabuta