ROS index logo ROS Index
  1. Home
  2. Packages
  3. autoware_sensing_launch
 
No version for distro humble showing github. Known supported distros are highlighted in the buttons above.
Package symbol

autoware_sensing_launch package from autoware_launch repo

autoware_launch autoware_sample_designs autoware_control_config autoware_control_launch autoware_localization_config autoware_localization_launch autoware_map_config autoware_map_launch autoware_perception_config autoware_perception_launch autoware_planning_config autoware_planning_launch autoware_sensing_launch autoware_sensor_driver_launch awsim_labs_sensor_kit_description awsim_labs_sensor_kit_launch awsim_sensor_kit_description awsim_sensor_kit_launch carla_sensor_kit_description carla_sensor_kit_launch sample_sensor_kit_description sample_sensor_kit_launch single_lidar_sensor_kit_description single_lidar_sensor_kit_launch tier4_autoware_api_launch tier4_simulator_launch tier4_system_launch tier4_vehicle_launch awsim_labs_vehicle_description awsim_labs_vehicle_launch sample_vehicle_description sample_vehicle_launch

ROS Distro
github

API Docs
Browse Code

Package Summary

Version 0.53.0
License Apache License 2.0
Build type AMENT_CMAKE
Use RECOMMENDED

Repository Summary

Description
Checkout URI https://github.com/autowarefoundation/autoware_launch.git
VCS Type git
VCS Version main
Last Updated 2026-10-26
Dev Status UNKNOWN
Released UNRELEASED
Contributing Help Wanted (-)
Good First Issues (-)
Pull Requests to Review (-)

Package Description

Sensor data pre-processing pipelines shared by sensor kits

Maintainers

  • David Wong
  • Max Schmeller
  • Dai Nguyen

Authors

No additional authors.
autoware_universe_launch/autoware_sensing_launch/README.md

autoware_sensing_launch

Shared sensing launch library. Sensor kits own the sensing entry point (<sensor_model>_launch/launch/sensing.launch.xml, included by component_sensing.launch.xml); this package hosts the reusable pre/post-processing units the kits build on. Driver launch files live in the driver layer package autoware_sensor_driver_launch.

Container ownership and addressing

Sensor kits create the containers; units only load nodes into them. pointcloud_container.launch.xml is the one unit that creates a container, and it loads nothing into it.

Units that load composable nodes take the target container as an absolute path (target_container) supplied by the caller; units never derive it from their own namespace. Callers construct the path from their literal namespace knowledge and must keep it in step with their namespace pushes. The one exception is the concatenation units, whose pointcloud_container_name keeps its historical name and receives the top-level pointcloud container path passed through unchanged.

Launch files

  • pointcloud_container.launch.xml — component container for one sensor unit, created at container_namespace/container_name; use_multithread selects component_container_mt.
  • pointcloud_crop_boxes.launch.xml — vehicle body crop box followed by a secondary crop box, loaded into an existing container. Used on its own where the remaining preprocessing stages do not apply.
  • pointcloud_preprocessor_cpu.launch.xml — per-lidar preprocessing chain: the crop box unit, then distortion correction and ring outlier filter, loaded into an existing container. The secondary crop box is named by secondary_crop_box (e.g. mirror, wheels).
  • pointcloud_preprocessor_cuda.launch.xml — the same chain folded into one CUDA node; crop box stages come as folded lists in crop_box_param_path.
  • pointcloud_concatenation_cpu.launch.xml, pointcloud_concatenation_cuda.launch.xml — multi-lidar concatenation with time synchronization; the kit passes its concatenate_and_time_sync_node_param_path (input topic list is kit-specific).
  • blockage_diagnostics.launch.xml — blockage diagnostics on the raw pointcloud; angle_range defaults to the configured device angles (float-formatted).
  • gnss_postprocessor.launch.xml — septentrio heading converter plus NavSatFix-to-MGRS pose (autoware_gnss_poser).
  • imu_corrector.launch.xml — imu corrector plus gyro bias estimator (autoware_imu_corrector).
  • radar_tracks_postprocessor.launch.xml — radar tracks noise filter plus message converter.

Config

  • distortion_corrector_node.param.yaml, ring_outlier_filter_node.param.yaml — shared preprocessing defaults; kits with different sensors or thresholds pass their own file via the *_param_path arguments (the outlier shape parameters such as vertical_bins only take effect when publish_outlier_pointcloud is enabled).
  • blockage_diagnostics.param.yaml — blockage diagnostics defaults; runtime shape parameters (angle range, bins) are set per sensor via launch arguments.
  • radar_tracks_msgs_converter.param.yaml, radar_tracks_noise_filter.param.yaml — radar postprocessing defaults.

Crop box param files are owned by the sensor kits (they encode vehicle geometry). Kits generate them offline with:

scripts/generate_crop_box_params.py <vehicle_info.param.yaml> <mirror.param.yaml> <output_dir>

and pass the outputs via the crop_box_filter_{self,secondary}_param_path arguments.

Scripts

  • septentrio_heading_converter.py — runtime node converting the septentrio AttEuler heading to a GnssInsOrientationStamped (installed).
  • generate_crop_box_params.py — offline generator for kit crop box files (not installed).

Package Dependencies

Please see <exec_depend> in package.xml.

CHANGELOG

Changelog for package autoware_sensing_launch

0.53.0 (2026-10-13)

  • Merge remote-tracking branch 'origin/main' into tmp/bot/bump_version_base

  • feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled (#1987)

    * feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled autoware_universe#13074 moves BlockageDiagComponent to agnocast_wrapper::Node, which is not designed to run inside a component container. Start the node as a standalone process when ENABLE_AGNOCAST=1 and keep loading it into target_container otherwise, as component_state_monitor.launch.py does for topic_state_monitor. The node name, namespace, remaps and parameters are identical in both branches.

    * refactor(autoware_sensing_launch): launch blockage_diag through the autoware_node action ---------

  • feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled (#1990)

    * feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled autoware_universe#13384 moves CropBoxFilterComponent to agnocast_wrapper::Node on an agnocast-only executor, which the component containers do not set up. Start each of the eight crop_box_filter nodes as a standalone process when ENABLE_AGNOCAST=1 and keep loading them into their container otherwise. The node names, namespaces, remaps and parameters are the same in both branches.

    * refactor(autoware_launch): launch crop_box_filter through the autoware_node action ---------

  • refactor(sensing)!: separate sensor driver and pre-processing layers from sensor kit configuration (#1938)

    * refactor(launch)!: rename tier4_sensing_launch to autoware_sensing_launch tier4_sensing_launch held one 19-line pass-through page that pushed the sensing namespace and included the sensor kit entry. The component page now includes that entry directly and pushes the namespace itself, so the layer is gone and the package is left as the home of the shared sensing units added next.

    • move the package to autoware_universe_launch/autoware_sensing_launch

    - rename tier4_sensing_component.launch.xml to components/component_sensing.launch.xml

    - drop the pass-through page and its diagram Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensing_launch): add the shared sensing unit library Declarative XML units for the sensor data pre/post-processing stages, each addressed by an absolute container path so a caller can place it in any container it owns.

    • pointcloud_container, the only unit that creates a container

    - pointcloud_crop_boxes, pointcloud_preprocessor_{cpu,cuda}, pointcloud_concatenation_{cpu,cuda}

    - blockage_diagnostics, gnss_postprocessor, imu_corrector, radar_tracks_postprocessor

    • shared preprocessing, blockage and radar defaults under config/

    - generate_crop_box_params.py, which turns vehicle info and mirror geometry into the crop box files the kits commit and pass in; septentrio_heading_converter.py is installed as a runtime node The parameter files and the heading converter are ported from tier4/aip_launcher (aip_common_sensor_launch), Apache-2.0 under the same TIER IV copyright. Nothing includes these units yet. Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensor_driver_launch): add the sensor driver layer The hardware interface layer, kept apart from the pre-processing units so that a sensor make and model can be brought into the ROS graph on its own.

    - nebula_{velodyne,robosense,hesai}: per-make driver chains, configured by the model param file shipped in the nebula driver package

    - velodyne_{VLP16,VLS128}, robosense_{Bpearl,Helios}, hesai_{XT32,OT128}: per-model wrappers

    - every file loads its node into target_container, an existing container owned by the caller; this package never creates one The

File truncated at 100 lines see the full file

Launch files

  • launch/blockage_diagnostics.launch.xml
      • target_container
      • node_name [default: blockage_diag]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • output/pointcloud [default: blockage_diag/pointcloud]
      • cloud_min_angle [default: 0.0]
      • cloud_max_angle [default: 360.0]
      • angle_range [default: [$(var cloud_min_angle), $(var cloud_max_angle)]]
      • max_range [default: 300.0]
      • horizontal_ring_id [default: 64]
      • vertical_bins [default: 128]
      • is_channel_order_top2down [default: true]
      • horizontal_resolution [default: 0.4]
      • blockage_diagnostics_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/blockage_diagnostics.param.yaml]
  • launch/gnss_postprocessor.launch.xml
      • input/fix [default: septentrio/nav_sat_fix]
      • input/atteuler [default: septentrio/atteuler]
      • input/orientation [default: septentrio/orientation]
      • output/pose [default: pose]
      • output/pose_with_covariance [default: pose_with_covariance]
      • output/fixed [default: fixed]
  • launch/imu_corrector.launch.xml
      • input/imu_raw
      • output/imu_data [default: imu_data]
      • input/odometry [default: /localization/kinematic_state]
      • input/pose_with_covariance [default: /localization/pose_estimator/pose_with_covariance]
      • output/imu_scaled [default: /sensing/imu_output_scaled]
      • imu_corrector_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/imu_corrector.param.yaml]
      • gyro_scale_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_scale_estimator.param.yaml]
      • gyro_bias_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_bias_estimator.param.yaml]
  • launch/pointcloud_concatenation_cpu.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • use_intra_process [default: false]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_concatenation_cuda.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_container.launch.xml
      • container_namespace
      • container_name [default: pointcloud_container]
      • use_multithread [default: false]
  • launch/pointcloud_crop_boxes.launch.xml
      • target_container
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • intermediate/pointcloud [default: self_cropped/pointcloud_ex]
      • output/pointcloud [default: $(var secondary_crop_box)_cropped/pointcloud_ex]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
  • launch/pointcloud_preprocessor_cpu.launch.xml
      • target_container
      • frame_id [default: lidar]
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • output_as_sensor_frame [default: true]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/pointcloud_preprocessor_cuda.launch.xml
      • target_container
      • enable_ring_outlier_filter [default: true]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/radar_tracks_postprocessor.launch.xml
      • input/tracks [default: objects_raw]
      • input/odometry [default: /localization/kinematic_state]
      • output/noise_tracks [default: noise_objects]
      • output/filtered_tracks [default: filtered_objects]
      • output/radar_detected_objects [default: detected_objects]
      • output/radar_tracked_objects [default: tracked_objects]
      • radar_tracks_msgs_converter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_msgs_converter.param.yaml]
      • radar_tracks_noise_filter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_noise_filter.param.yaml]

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged autoware_sensing_launch at Robotics Stack Exchange

No version for distro jazzy showing github. Known supported distros are highlighted in the buttons above.
Package symbol

autoware_sensing_launch package from autoware_launch repo

autoware_launch autoware_sample_designs autoware_control_config autoware_control_launch autoware_localization_config autoware_localization_launch autoware_map_config autoware_map_launch autoware_perception_config autoware_perception_launch autoware_planning_config autoware_planning_launch autoware_sensing_launch autoware_sensor_driver_launch awsim_labs_sensor_kit_description awsim_labs_sensor_kit_launch awsim_sensor_kit_description awsim_sensor_kit_launch carla_sensor_kit_description carla_sensor_kit_launch sample_sensor_kit_description sample_sensor_kit_launch single_lidar_sensor_kit_description single_lidar_sensor_kit_launch tier4_autoware_api_launch tier4_simulator_launch tier4_system_launch tier4_vehicle_launch awsim_labs_vehicle_description awsim_labs_vehicle_launch sample_vehicle_description sample_vehicle_launch

ROS Distro
github

API Docs
Browse Code

Package Summary

Version 0.53.0
License Apache License 2.0
Build type AMENT_CMAKE
Use RECOMMENDED

Repository Summary

Description
Checkout URI https://github.com/autowarefoundation/autoware_launch.git
VCS Type git
VCS Version main
Last Updated 2026-10-26
Dev Status UNKNOWN
Released UNRELEASED
Contributing Help Wanted (-)
Good First Issues (-)
Pull Requests to Review (-)

Package Description

Sensor data pre-processing pipelines shared by sensor kits

Maintainers

  • David Wong
  • Max Schmeller
  • Dai Nguyen

Authors

No additional authors.
autoware_universe_launch/autoware_sensing_launch/README.md

autoware_sensing_launch

Shared sensing launch library. Sensor kits own the sensing entry point (<sensor_model>_launch/launch/sensing.launch.xml, included by component_sensing.launch.xml); this package hosts the reusable pre/post-processing units the kits build on. Driver launch files live in the driver layer package autoware_sensor_driver_launch.

Container ownership and addressing

Sensor kits create the containers; units only load nodes into them. pointcloud_container.launch.xml is the one unit that creates a container, and it loads nothing into it.

Units that load composable nodes take the target container as an absolute path (target_container) supplied by the caller; units never derive it from their own namespace. Callers construct the path from their literal namespace knowledge and must keep it in step with their namespace pushes. The one exception is the concatenation units, whose pointcloud_container_name keeps its historical name and receives the top-level pointcloud container path passed through unchanged.

Launch files

  • pointcloud_container.launch.xml — component container for one sensor unit, created at container_namespace/container_name; use_multithread selects component_container_mt.
  • pointcloud_crop_boxes.launch.xml — vehicle body crop box followed by a secondary crop box, loaded into an existing container. Used on its own where the remaining preprocessing stages do not apply.
  • pointcloud_preprocessor_cpu.launch.xml — per-lidar preprocessing chain: the crop box unit, then distortion correction and ring outlier filter, loaded into an existing container. The secondary crop box is named by secondary_crop_box (e.g. mirror, wheels).
  • pointcloud_preprocessor_cuda.launch.xml — the same chain folded into one CUDA node; crop box stages come as folded lists in crop_box_param_path.
  • pointcloud_concatenation_cpu.launch.xml, pointcloud_concatenation_cuda.launch.xml — multi-lidar concatenation with time synchronization; the kit passes its concatenate_and_time_sync_node_param_path (input topic list is kit-specific).
  • blockage_diagnostics.launch.xml — blockage diagnostics on the raw pointcloud; angle_range defaults to the configured device angles (float-formatted).
  • gnss_postprocessor.launch.xml — septentrio heading converter plus NavSatFix-to-MGRS pose (autoware_gnss_poser).
  • imu_corrector.launch.xml — imu corrector plus gyro bias estimator (autoware_imu_corrector).
  • radar_tracks_postprocessor.launch.xml — radar tracks noise filter plus message converter.

Config

  • distortion_corrector_node.param.yaml, ring_outlier_filter_node.param.yaml — shared preprocessing defaults; kits with different sensors or thresholds pass their own file via the *_param_path arguments (the outlier shape parameters such as vertical_bins only take effect when publish_outlier_pointcloud is enabled).
  • blockage_diagnostics.param.yaml — blockage diagnostics defaults; runtime shape parameters (angle range, bins) are set per sensor via launch arguments.
  • radar_tracks_msgs_converter.param.yaml, radar_tracks_noise_filter.param.yaml — radar postprocessing defaults.

Crop box param files are owned by the sensor kits (they encode vehicle geometry). Kits generate them offline with:

scripts/generate_crop_box_params.py <vehicle_info.param.yaml> <mirror.param.yaml> <output_dir>

and pass the outputs via the crop_box_filter_{self,secondary}_param_path arguments.

Scripts

  • septentrio_heading_converter.py — runtime node converting the septentrio AttEuler heading to a GnssInsOrientationStamped (installed).
  • generate_crop_box_params.py — offline generator for kit crop box files (not installed).

Package Dependencies

Please see <exec_depend> in package.xml.

CHANGELOG

Changelog for package autoware_sensing_launch

0.53.0 (2026-10-13)

  • Merge remote-tracking branch 'origin/main' into tmp/bot/bump_version_base

  • feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled (#1987)

    * feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled autoware_universe#13074 moves BlockageDiagComponent to agnocast_wrapper::Node, which is not designed to run inside a component container. Start the node as a standalone process when ENABLE_AGNOCAST=1 and keep loading it into target_container otherwise, as component_state_monitor.launch.py does for topic_state_monitor. The node name, namespace, remaps and parameters are identical in both branches.

    * refactor(autoware_sensing_launch): launch blockage_diag through the autoware_node action ---------

  • feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled (#1990)

    * feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled autoware_universe#13384 moves CropBoxFilterComponent to agnocast_wrapper::Node on an agnocast-only executor, which the component containers do not set up. Start each of the eight crop_box_filter nodes as a standalone process when ENABLE_AGNOCAST=1 and keep loading them into their container otherwise. The node names, namespaces, remaps and parameters are the same in both branches.

    * refactor(autoware_launch): launch crop_box_filter through the autoware_node action ---------

  • refactor(sensing)!: separate sensor driver and pre-processing layers from sensor kit configuration (#1938)

    * refactor(launch)!: rename tier4_sensing_launch to autoware_sensing_launch tier4_sensing_launch held one 19-line pass-through page that pushed the sensing namespace and included the sensor kit entry. The component page now includes that entry directly and pushes the namespace itself, so the layer is gone and the package is left as the home of the shared sensing units added next.

    • move the package to autoware_universe_launch/autoware_sensing_launch

    - rename tier4_sensing_component.launch.xml to components/component_sensing.launch.xml

    - drop the pass-through page and its diagram Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensing_launch): add the shared sensing unit library Declarative XML units for the sensor data pre/post-processing stages, each addressed by an absolute container path so a caller can place it in any container it owns.

    • pointcloud_container, the only unit that creates a container

    - pointcloud_crop_boxes, pointcloud_preprocessor_{cpu,cuda}, pointcloud_concatenation_{cpu,cuda}

    - blockage_diagnostics, gnss_postprocessor, imu_corrector, radar_tracks_postprocessor

    • shared preprocessing, blockage and radar defaults under config/

    - generate_crop_box_params.py, which turns vehicle info and mirror geometry into the crop box files the kits commit and pass in; septentrio_heading_converter.py is installed as a runtime node The parameter files and the heading converter are ported from tier4/aip_launcher (aip_common_sensor_launch), Apache-2.0 under the same TIER IV copyright. Nothing includes these units yet. Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensor_driver_launch): add the sensor driver layer The hardware interface layer, kept apart from the pre-processing units so that a sensor make and model can be brought into the ROS graph on its own.

    - nebula_{velodyne,robosense,hesai}: per-make driver chains, configured by the model param file shipped in the nebula driver package

    - velodyne_{VLP16,VLS128}, robosense_{Bpearl,Helios}, hesai_{XT32,OT128}: per-model wrappers

    - every file loads its node into target_container, an existing container owned by the caller; this package never creates one The

File truncated at 100 lines see the full file

Launch files

  • launch/blockage_diagnostics.launch.xml
      • target_container
      • node_name [default: blockage_diag]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • output/pointcloud [default: blockage_diag/pointcloud]
      • cloud_min_angle [default: 0.0]
      • cloud_max_angle [default: 360.0]
      • angle_range [default: [$(var cloud_min_angle), $(var cloud_max_angle)]]
      • max_range [default: 300.0]
      • horizontal_ring_id [default: 64]
      • vertical_bins [default: 128]
      • is_channel_order_top2down [default: true]
      • horizontal_resolution [default: 0.4]
      • blockage_diagnostics_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/blockage_diagnostics.param.yaml]
  • launch/gnss_postprocessor.launch.xml
      • input/fix [default: septentrio/nav_sat_fix]
      • input/atteuler [default: septentrio/atteuler]
      • input/orientation [default: septentrio/orientation]
      • output/pose [default: pose]
      • output/pose_with_covariance [default: pose_with_covariance]
      • output/fixed [default: fixed]
  • launch/imu_corrector.launch.xml
      • input/imu_raw
      • output/imu_data [default: imu_data]
      • input/odometry [default: /localization/kinematic_state]
      • input/pose_with_covariance [default: /localization/pose_estimator/pose_with_covariance]
      • output/imu_scaled [default: /sensing/imu_output_scaled]
      • imu_corrector_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/imu_corrector.param.yaml]
      • gyro_scale_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_scale_estimator.param.yaml]
      • gyro_bias_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_bias_estimator.param.yaml]
  • launch/pointcloud_concatenation_cpu.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • use_intra_process [default: false]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_concatenation_cuda.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_container.launch.xml
      • container_namespace
      • container_name [default: pointcloud_container]
      • use_multithread [default: false]
  • launch/pointcloud_crop_boxes.launch.xml
      • target_container
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • intermediate/pointcloud [default: self_cropped/pointcloud_ex]
      • output/pointcloud [default: $(var secondary_crop_box)_cropped/pointcloud_ex]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
  • launch/pointcloud_preprocessor_cpu.launch.xml
      • target_container
      • frame_id [default: lidar]
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • output_as_sensor_frame [default: true]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/pointcloud_preprocessor_cuda.launch.xml
      • target_container
      • enable_ring_outlier_filter [default: true]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/radar_tracks_postprocessor.launch.xml
      • input/tracks [default: objects_raw]
      • input/odometry [default: /localization/kinematic_state]
      • output/noise_tracks [default: noise_objects]
      • output/filtered_tracks [default: filtered_objects]
      • output/radar_detected_objects [default: detected_objects]
      • output/radar_tracked_objects [default: tracked_objects]
      • radar_tracks_msgs_converter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_msgs_converter.param.yaml]
      • radar_tracks_noise_filter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_noise_filter.param.yaml]

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged autoware_sensing_launch at Robotics Stack Exchange

No version for distro kilted showing github. Known supported distros are highlighted in the buttons above.
Package symbol

autoware_sensing_launch package from autoware_launch repo

autoware_launch autoware_sample_designs autoware_control_config autoware_control_launch autoware_localization_config autoware_localization_launch autoware_map_config autoware_map_launch autoware_perception_config autoware_perception_launch autoware_planning_config autoware_planning_launch autoware_sensing_launch autoware_sensor_driver_launch awsim_labs_sensor_kit_description awsim_labs_sensor_kit_launch awsim_sensor_kit_description awsim_sensor_kit_launch carla_sensor_kit_description carla_sensor_kit_launch sample_sensor_kit_description sample_sensor_kit_launch single_lidar_sensor_kit_description single_lidar_sensor_kit_launch tier4_autoware_api_launch tier4_simulator_launch tier4_system_launch tier4_vehicle_launch awsim_labs_vehicle_description awsim_labs_vehicle_launch sample_vehicle_description sample_vehicle_launch

ROS Distro
github

API Docs
Browse Code

Package Summary

Version 0.53.0
License Apache License 2.0
Build type AMENT_CMAKE
Use RECOMMENDED

Repository Summary

Description
Checkout URI https://github.com/autowarefoundation/autoware_launch.git
VCS Type git
VCS Version main
Last Updated 2026-10-26
Dev Status UNKNOWN
Released UNRELEASED
Contributing Help Wanted (-)
Good First Issues (-)
Pull Requests to Review (-)

Package Description

Sensor data pre-processing pipelines shared by sensor kits

Maintainers

  • David Wong
  • Max Schmeller
  • Dai Nguyen

Authors

No additional authors.
autoware_universe_launch/autoware_sensing_launch/README.md

autoware_sensing_launch

Shared sensing launch library. Sensor kits own the sensing entry point (<sensor_model>_launch/launch/sensing.launch.xml, included by component_sensing.launch.xml); this package hosts the reusable pre/post-processing units the kits build on. Driver launch files live in the driver layer package autoware_sensor_driver_launch.

Container ownership and addressing

Sensor kits create the containers; units only load nodes into them. pointcloud_container.launch.xml is the one unit that creates a container, and it loads nothing into it.

Units that load composable nodes take the target container as an absolute path (target_container) supplied by the caller; units never derive it from their own namespace. Callers construct the path from their literal namespace knowledge and must keep it in step with their namespace pushes. The one exception is the concatenation units, whose pointcloud_container_name keeps its historical name and receives the top-level pointcloud container path passed through unchanged.

Launch files

  • pointcloud_container.launch.xml — component container for one sensor unit, created at container_namespace/container_name; use_multithread selects component_container_mt.
  • pointcloud_crop_boxes.launch.xml — vehicle body crop box followed by a secondary crop box, loaded into an existing container. Used on its own where the remaining preprocessing stages do not apply.
  • pointcloud_preprocessor_cpu.launch.xml — per-lidar preprocessing chain: the crop box unit, then distortion correction and ring outlier filter, loaded into an existing container. The secondary crop box is named by secondary_crop_box (e.g. mirror, wheels).
  • pointcloud_preprocessor_cuda.launch.xml — the same chain folded into one CUDA node; crop box stages come as folded lists in crop_box_param_path.
  • pointcloud_concatenation_cpu.launch.xml, pointcloud_concatenation_cuda.launch.xml — multi-lidar concatenation with time synchronization; the kit passes its concatenate_and_time_sync_node_param_path (input topic list is kit-specific).
  • blockage_diagnostics.launch.xml — blockage diagnostics on the raw pointcloud; angle_range defaults to the configured device angles (float-formatted).
  • gnss_postprocessor.launch.xml — septentrio heading converter plus NavSatFix-to-MGRS pose (autoware_gnss_poser).
  • imu_corrector.launch.xml — imu corrector plus gyro bias estimator (autoware_imu_corrector).
  • radar_tracks_postprocessor.launch.xml — radar tracks noise filter plus message converter.

Config

  • distortion_corrector_node.param.yaml, ring_outlier_filter_node.param.yaml — shared preprocessing defaults; kits with different sensors or thresholds pass their own file via the *_param_path arguments (the outlier shape parameters such as vertical_bins only take effect when publish_outlier_pointcloud is enabled).
  • blockage_diagnostics.param.yaml — blockage diagnostics defaults; runtime shape parameters (angle range, bins) are set per sensor via launch arguments.
  • radar_tracks_msgs_converter.param.yaml, radar_tracks_noise_filter.param.yaml — radar postprocessing defaults.

Crop box param files are owned by the sensor kits (they encode vehicle geometry). Kits generate them offline with:

scripts/generate_crop_box_params.py <vehicle_info.param.yaml> <mirror.param.yaml> <output_dir>

and pass the outputs via the crop_box_filter_{self,secondary}_param_path arguments.

Scripts

  • septentrio_heading_converter.py — runtime node converting the septentrio AttEuler heading to a GnssInsOrientationStamped (installed).
  • generate_crop_box_params.py — offline generator for kit crop box files (not installed).

Package Dependencies

Please see <exec_depend> in package.xml.

CHANGELOG

Changelog for package autoware_sensing_launch

0.53.0 (2026-10-13)

  • Merge remote-tracking branch 'origin/main' into tmp/bot/bump_version_base

  • feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled (#1987)

    * feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled autoware_universe#13074 moves BlockageDiagComponent to agnocast_wrapper::Node, which is not designed to run inside a component container. Start the node as a standalone process when ENABLE_AGNOCAST=1 and keep loading it into target_container otherwise, as component_state_monitor.launch.py does for topic_state_monitor. The node name, namespace, remaps and parameters are identical in both branches.

    * refactor(autoware_sensing_launch): launch blockage_diag through the autoware_node action ---------

  • feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled (#1990)

    * feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled autoware_universe#13384 moves CropBoxFilterComponent to agnocast_wrapper::Node on an agnocast-only executor, which the component containers do not set up. Start each of the eight crop_box_filter nodes as a standalone process when ENABLE_AGNOCAST=1 and keep loading them into their container otherwise. The node names, namespaces, remaps and parameters are the same in both branches.

    * refactor(autoware_launch): launch crop_box_filter through the autoware_node action ---------

  • refactor(sensing)!: separate sensor driver and pre-processing layers from sensor kit configuration (#1938)

    * refactor(launch)!: rename tier4_sensing_launch to autoware_sensing_launch tier4_sensing_launch held one 19-line pass-through page that pushed the sensing namespace and included the sensor kit entry. The component page now includes that entry directly and pushes the namespace itself, so the layer is gone and the package is left as the home of the shared sensing units added next.

    • move the package to autoware_universe_launch/autoware_sensing_launch

    - rename tier4_sensing_component.launch.xml to components/component_sensing.launch.xml

    - drop the pass-through page and its diagram Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensing_launch): add the shared sensing unit library Declarative XML units for the sensor data pre/post-processing stages, each addressed by an absolute container path so a caller can place it in any container it owns.

    • pointcloud_container, the only unit that creates a container

    - pointcloud_crop_boxes, pointcloud_preprocessor_{cpu,cuda}, pointcloud_concatenation_{cpu,cuda}

    - blockage_diagnostics, gnss_postprocessor, imu_corrector, radar_tracks_postprocessor

    • shared preprocessing, blockage and radar defaults under config/

    - generate_crop_box_params.py, which turns vehicle info and mirror geometry into the crop box files the kits commit and pass in; septentrio_heading_converter.py is installed as a runtime node The parameter files and the heading converter are ported from tier4/aip_launcher (aip_common_sensor_launch), Apache-2.0 under the same TIER IV copyright. Nothing includes these units yet. Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensor_driver_launch): add the sensor driver layer The hardware interface layer, kept apart from the pre-processing units so that a sensor make and model can be brought into the ROS graph on its own.

    - nebula_{velodyne,robosense,hesai}: per-make driver chains, configured by the model param file shipped in the nebula driver package

    - velodyne_{VLP16,VLS128}, robosense_{Bpearl,Helios}, hesai_{XT32,OT128}: per-model wrappers

    - every file loads its node into target_container, an existing container owned by the caller; this package never creates one The

File truncated at 100 lines see the full file

Launch files

  • launch/blockage_diagnostics.launch.xml
      • target_container
      • node_name [default: blockage_diag]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • output/pointcloud [default: blockage_diag/pointcloud]
      • cloud_min_angle [default: 0.0]
      • cloud_max_angle [default: 360.0]
      • angle_range [default: [$(var cloud_min_angle), $(var cloud_max_angle)]]
      • max_range [default: 300.0]
      • horizontal_ring_id [default: 64]
      • vertical_bins [default: 128]
      • is_channel_order_top2down [default: true]
      • horizontal_resolution [default: 0.4]
      • blockage_diagnostics_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/blockage_diagnostics.param.yaml]
  • launch/gnss_postprocessor.launch.xml
      • input/fix [default: septentrio/nav_sat_fix]
      • input/atteuler [default: septentrio/atteuler]
      • input/orientation [default: septentrio/orientation]
      • output/pose [default: pose]
      • output/pose_with_covariance [default: pose_with_covariance]
      • output/fixed [default: fixed]
  • launch/imu_corrector.launch.xml
      • input/imu_raw
      • output/imu_data [default: imu_data]
      • input/odometry [default: /localization/kinematic_state]
      • input/pose_with_covariance [default: /localization/pose_estimator/pose_with_covariance]
      • output/imu_scaled [default: /sensing/imu_output_scaled]
      • imu_corrector_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/imu_corrector.param.yaml]
      • gyro_scale_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_scale_estimator.param.yaml]
      • gyro_bias_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_bias_estimator.param.yaml]
  • launch/pointcloud_concatenation_cpu.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • use_intra_process [default: false]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_concatenation_cuda.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_container.launch.xml
      • container_namespace
      • container_name [default: pointcloud_container]
      • use_multithread [default: false]
  • launch/pointcloud_crop_boxes.launch.xml
      • target_container
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • intermediate/pointcloud [default: self_cropped/pointcloud_ex]
      • output/pointcloud [default: $(var secondary_crop_box)_cropped/pointcloud_ex]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
  • launch/pointcloud_preprocessor_cpu.launch.xml
      • target_container
      • frame_id [default: lidar]
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • output_as_sensor_frame [default: true]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/pointcloud_preprocessor_cuda.launch.xml
      • target_container
      • enable_ring_outlier_filter [default: true]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/radar_tracks_postprocessor.launch.xml
      • input/tracks [default: objects_raw]
      • input/odometry [default: /localization/kinematic_state]
      • output/noise_tracks [default: noise_objects]
      • output/filtered_tracks [default: filtered_objects]
      • output/radar_detected_objects [default: detected_objects]
      • output/radar_tracked_objects [default: tracked_objects]
      • radar_tracks_msgs_converter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_msgs_converter.param.yaml]
      • radar_tracks_noise_filter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_noise_filter.param.yaml]

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged autoware_sensing_launch at Robotics Stack Exchange

No version for distro lyrical showing github. Known supported distros are highlighted in the buttons above.
Package symbol

autoware_sensing_launch package from autoware_launch repo

autoware_launch autoware_sample_designs autoware_control_config autoware_control_launch autoware_localization_config autoware_localization_launch autoware_map_config autoware_map_launch autoware_perception_config autoware_perception_launch autoware_planning_config autoware_planning_launch autoware_sensing_launch autoware_sensor_driver_launch awsim_labs_sensor_kit_description awsim_labs_sensor_kit_launch awsim_sensor_kit_description awsim_sensor_kit_launch carla_sensor_kit_description carla_sensor_kit_launch sample_sensor_kit_description sample_sensor_kit_launch single_lidar_sensor_kit_description single_lidar_sensor_kit_launch tier4_autoware_api_launch tier4_simulator_launch tier4_system_launch tier4_vehicle_launch awsim_labs_vehicle_description awsim_labs_vehicle_launch sample_vehicle_description sample_vehicle_launch

ROS Distro
github

API Docs
Browse Code

Package Summary

Version 0.53.0
License Apache License 2.0
Build type AMENT_CMAKE
Use RECOMMENDED

Repository Summary

Description
Checkout URI https://github.com/autowarefoundation/autoware_launch.git
VCS Type git
VCS Version main
Last Updated 2026-10-26
Dev Status UNKNOWN
Released UNRELEASED
Contributing Help Wanted (-)
Good First Issues (-)
Pull Requests to Review (-)

Package Description

Sensor data pre-processing pipelines shared by sensor kits

Maintainers

  • David Wong
  • Max Schmeller
  • Dai Nguyen

Authors

No additional authors.
autoware_universe_launch/autoware_sensing_launch/README.md

autoware_sensing_launch

Shared sensing launch library. Sensor kits own the sensing entry point (<sensor_model>_launch/launch/sensing.launch.xml, included by component_sensing.launch.xml); this package hosts the reusable pre/post-processing units the kits build on. Driver launch files live in the driver layer package autoware_sensor_driver_launch.

Container ownership and addressing

Sensor kits create the containers; units only load nodes into them. pointcloud_container.launch.xml is the one unit that creates a container, and it loads nothing into it.

Units that load composable nodes take the target container as an absolute path (target_container) supplied by the caller; units never derive it from their own namespace. Callers construct the path from their literal namespace knowledge and must keep it in step with their namespace pushes. The one exception is the concatenation units, whose pointcloud_container_name keeps its historical name and receives the top-level pointcloud container path passed through unchanged.

Launch files

  • pointcloud_container.launch.xml — component container for one sensor unit, created at container_namespace/container_name; use_multithread selects component_container_mt.
  • pointcloud_crop_boxes.launch.xml — vehicle body crop box followed by a secondary crop box, loaded into an existing container. Used on its own where the remaining preprocessing stages do not apply.
  • pointcloud_preprocessor_cpu.launch.xml — per-lidar preprocessing chain: the crop box unit, then distortion correction and ring outlier filter, loaded into an existing container. The secondary crop box is named by secondary_crop_box (e.g. mirror, wheels).
  • pointcloud_preprocessor_cuda.launch.xml — the same chain folded into one CUDA node; crop box stages come as folded lists in crop_box_param_path.
  • pointcloud_concatenation_cpu.launch.xml, pointcloud_concatenation_cuda.launch.xml — multi-lidar concatenation with time synchronization; the kit passes its concatenate_and_time_sync_node_param_path (input topic list is kit-specific).
  • blockage_diagnostics.launch.xml — blockage diagnostics on the raw pointcloud; angle_range defaults to the configured device angles (float-formatted).
  • gnss_postprocessor.launch.xml — septentrio heading converter plus NavSatFix-to-MGRS pose (autoware_gnss_poser).
  • imu_corrector.launch.xml — imu corrector plus gyro bias estimator (autoware_imu_corrector).
  • radar_tracks_postprocessor.launch.xml — radar tracks noise filter plus message converter.

Config

  • distortion_corrector_node.param.yaml, ring_outlier_filter_node.param.yaml — shared preprocessing defaults; kits with different sensors or thresholds pass their own file via the *_param_path arguments (the outlier shape parameters such as vertical_bins only take effect when publish_outlier_pointcloud is enabled).
  • blockage_diagnostics.param.yaml — blockage diagnostics defaults; runtime shape parameters (angle range, bins) are set per sensor via launch arguments.
  • radar_tracks_msgs_converter.param.yaml, radar_tracks_noise_filter.param.yaml — radar postprocessing defaults.

Crop box param files are owned by the sensor kits (they encode vehicle geometry). Kits generate them offline with:

scripts/generate_crop_box_params.py <vehicle_info.param.yaml> <mirror.param.yaml> <output_dir>

and pass the outputs via the crop_box_filter_{self,secondary}_param_path arguments.

Scripts

  • septentrio_heading_converter.py — runtime node converting the septentrio AttEuler heading to a GnssInsOrientationStamped (installed).
  • generate_crop_box_params.py — offline generator for kit crop box files (not installed).

Package Dependencies

Please see <exec_depend> in package.xml.

CHANGELOG

Changelog for package autoware_sensing_launch

0.53.0 (2026-10-13)

  • Merge remote-tracking branch 'origin/main' into tmp/bot/bump_version_base

  • feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled (#1987)

    * feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled autoware_universe#13074 moves BlockageDiagComponent to agnocast_wrapper::Node, which is not designed to run inside a component container. Start the node as a standalone process when ENABLE_AGNOCAST=1 and keep loading it into target_container otherwise, as component_state_monitor.launch.py does for topic_state_monitor. The node name, namespace, remaps and parameters are identical in both branches.

    * refactor(autoware_sensing_launch): launch blockage_diag through the autoware_node action ---------

  • feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled (#1990)

    * feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled autoware_universe#13384 moves CropBoxFilterComponent to agnocast_wrapper::Node on an agnocast-only executor, which the component containers do not set up. Start each of the eight crop_box_filter nodes as a standalone process when ENABLE_AGNOCAST=1 and keep loading them into their container otherwise. The node names, namespaces, remaps and parameters are the same in both branches.

    * refactor(autoware_launch): launch crop_box_filter through the autoware_node action ---------

  • refactor(sensing)!: separate sensor driver and pre-processing layers from sensor kit configuration (#1938)

    * refactor(launch)!: rename tier4_sensing_launch to autoware_sensing_launch tier4_sensing_launch held one 19-line pass-through page that pushed the sensing namespace and included the sensor kit entry. The component page now includes that entry directly and pushes the namespace itself, so the layer is gone and the package is left as the home of the shared sensing units added next.

    • move the package to autoware_universe_launch/autoware_sensing_launch

    - rename tier4_sensing_component.launch.xml to components/component_sensing.launch.xml

    - drop the pass-through page and its diagram Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensing_launch): add the shared sensing unit library Declarative XML units for the sensor data pre/post-processing stages, each addressed by an absolute container path so a caller can place it in any container it owns.

    • pointcloud_container, the only unit that creates a container

    - pointcloud_crop_boxes, pointcloud_preprocessor_{cpu,cuda}, pointcloud_concatenation_{cpu,cuda}

    - blockage_diagnostics, gnss_postprocessor, imu_corrector, radar_tracks_postprocessor

    • shared preprocessing, blockage and radar defaults under config/

    - generate_crop_box_params.py, which turns vehicle info and mirror geometry into the crop box files the kits commit and pass in; septentrio_heading_converter.py is installed as a runtime node The parameter files and the heading converter are ported from tier4/aip_launcher (aip_common_sensor_launch), Apache-2.0 under the same TIER IV copyright. Nothing includes these units yet. Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensor_driver_launch): add the sensor driver layer The hardware interface layer, kept apart from the pre-processing units so that a sensor make and model can be brought into the ROS graph on its own.

    - nebula_{velodyne,robosense,hesai}: per-make driver chains, configured by the model param file shipped in the nebula driver package

    - velodyne_{VLP16,VLS128}, robosense_{Bpearl,Helios}, hesai_{XT32,OT128}: per-model wrappers

    - every file loads its node into target_container, an existing container owned by the caller; this package never creates one The

File truncated at 100 lines see the full file

Launch files

  • launch/blockage_diagnostics.launch.xml
      • target_container
      • node_name [default: blockage_diag]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • output/pointcloud [default: blockage_diag/pointcloud]
      • cloud_min_angle [default: 0.0]
      • cloud_max_angle [default: 360.0]
      • angle_range [default: [$(var cloud_min_angle), $(var cloud_max_angle)]]
      • max_range [default: 300.0]
      • horizontal_ring_id [default: 64]
      • vertical_bins [default: 128]
      • is_channel_order_top2down [default: true]
      • horizontal_resolution [default: 0.4]
      • blockage_diagnostics_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/blockage_diagnostics.param.yaml]
  • launch/gnss_postprocessor.launch.xml
      • input/fix [default: septentrio/nav_sat_fix]
      • input/atteuler [default: septentrio/atteuler]
      • input/orientation [default: septentrio/orientation]
      • output/pose [default: pose]
      • output/pose_with_covariance [default: pose_with_covariance]
      • output/fixed [default: fixed]
  • launch/imu_corrector.launch.xml
      • input/imu_raw
      • output/imu_data [default: imu_data]
      • input/odometry [default: /localization/kinematic_state]
      • input/pose_with_covariance [default: /localization/pose_estimator/pose_with_covariance]
      • output/imu_scaled [default: /sensing/imu_output_scaled]
      • imu_corrector_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/imu_corrector.param.yaml]
      • gyro_scale_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_scale_estimator.param.yaml]
      • gyro_bias_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_bias_estimator.param.yaml]
  • launch/pointcloud_concatenation_cpu.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • use_intra_process [default: false]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_concatenation_cuda.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_container.launch.xml
      • container_namespace
      • container_name [default: pointcloud_container]
      • use_multithread [default: false]
  • launch/pointcloud_crop_boxes.launch.xml
      • target_container
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • intermediate/pointcloud [default: self_cropped/pointcloud_ex]
      • output/pointcloud [default: $(var secondary_crop_box)_cropped/pointcloud_ex]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
  • launch/pointcloud_preprocessor_cpu.launch.xml
      • target_container
      • frame_id [default: lidar]
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • output_as_sensor_frame [default: true]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/pointcloud_preprocessor_cuda.launch.xml
      • target_container
      • enable_ring_outlier_filter [default: true]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/radar_tracks_postprocessor.launch.xml
      • input/tracks [default: objects_raw]
      • input/odometry [default: /localization/kinematic_state]
      • output/noise_tracks [default: noise_objects]
      • output/filtered_tracks [default: filtered_objects]
      • output/radar_detected_objects [default: detected_objects]
      • output/radar_tracked_objects [default: tracked_objects]
      • radar_tracks_msgs_converter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_msgs_converter.param.yaml]
      • radar_tracks_noise_filter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_noise_filter.param.yaml]

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged autoware_sensing_launch at Robotics Stack Exchange

No version for distro rolling showing github. Known supported distros are highlighted in the buttons above.
Package symbol

autoware_sensing_launch package from autoware_launch repo

autoware_launch autoware_sample_designs autoware_control_config autoware_control_launch autoware_localization_config autoware_localization_launch autoware_map_config autoware_map_launch autoware_perception_config autoware_perception_launch autoware_planning_config autoware_planning_launch autoware_sensing_launch autoware_sensor_driver_launch awsim_labs_sensor_kit_description awsim_labs_sensor_kit_launch awsim_sensor_kit_description awsim_sensor_kit_launch carla_sensor_kit_description carla_sensor_kit_launch sample_sensor_kit_description sample_sensor_kit_launch single_lidar_sensor_kit_description single_lidar_sensor_kit_launch tier4_autoware_api_launch tier4_simulator_launch tier4_system_launch tier4_vehicle_launch awsim_labs_vehicle_description awsim_labs_vehicle_launch sample_vehicle_description sample_vehicle_launch

ROS Distro
github

API Docs
Browse Code

Package Summary

Version 0.53.0
License Apache License 2.0
Build type AMENT_CMAKE
Use RECOMMENDED

Repository Summary

Description
Checkout URI https://github.com/autowarefoundation/autoware_launch.git
VCS Type git
VCS Version main
Last Updated 2026-10-26
Dev Status UNKNOWN
Released UNRELEASED
Contributing Help Wanted (-)
Good First Issues (-)
Pull Requests to Review (-)

Package Description

Sensor data pre-processing pipelines shared by sensor kits

Maintainers

  • David Wong
  • Max Schmeller
  • Dai Nguyen

Authors

No additional authors.
autoware_universe_launch/autoware_sensing_launch/README.md

autoware_sensing_launch

Shared sensing launch library. Sensor kits own the sensing entry point (<sensor_model>_launch/launch/sensing.launch.xml, included by component_sensing.launch.xml); this package hosts the reusable pre/post-processing units the kits build on. Driver launch files live in the driver layer package autoware_sensor_driver_launch.

Container ownership and addressing

Sensor kits create the containers; units only load nodes into them. pointcloud_container.launch.xml is the one unit that creates a container, and it loads nothing into it.

Units that load composable nodes take the target container as an absolute path (target_container) supplied by the caller; units never derive it from their own namespace. Callers construct the path from their literal namespace knowledge and must keep it in step with their namespace pushes. The one exception is the concatenation units, whose pointcloud_container_name keeps its historical name and receives the top-level pointcloud container path passed through unchanged.

Launch files

  • pointcloud_container.launch.xml — component container for one sensor unit, created at container_namespace/container_name; use_multithread selects component_container_mt.
  • pointcloud_crop_boxes.launch.xml — vehicle body crop box followed by a secondary crop box, loaded into an existing container. Used on its own where the remaining preprocessing stages do not apply.
  • pointcloud_preprocessor_cpu.launch.xml — per-lidar preprocessing chain: the crop box unit, then distortion correction and ring outlier filter, loaded into an existing container. The secondary crop box is named by secondary_crop_box (e.g. mirror, wheels).
  • pointcloud_preprocessor_cuda.launch.xml — the same chain folded into one CUDA node; crop box stages come as folded lists in crop_box_param_path.
  • pointcloud_concatenation_cpu.launch.xml, pointcloud_concatenation_cuda.launch.xml — multi-lidar concatenation with time synchronization; the kit passes its concatenate_and_time_sync_node_param_path (input topic list is kit-specific).
  • blockage_diagnostics.launch.xml — blockage diagnostics on the raw pointcloud; angle_range defaults to the configured device angles (float-formatted).
  • gnss_postprocessor.launch.xml — septentrio heading converter plus NavSatFix-to-MGRS pose (autoware_gnss_poser).
  • imu_corrector.launch.xml — imu corrector plus gyro bias estimator (autoware_imu_corrector).
  • radar_tracks_postprocessor.launch.xml — radar tracks noise filter plus message converter.

Config

  • distortion_corrector_node.param.yaml, ring_outlier_filter_node.param.yaml — shared preprocessing defaults; kits with different sensors or thresholds pass their own file via the *_param_path arguments (the outlier shape parameters such as vertical_bins only take effect when publish_outlier_pointcloud is enabled).
  • blockage_diagnostics.param.yaml — blockage diagnostics defaults; runtime shape parameters (angle range, bins) are set per sensor via launch arguments.
  • radar_tracks_msgs_converter.param.yaml, radar_tracks_noise_filter.param.yaml — radar postprocessing defaults.

Crop box param files are owned by the sensor kits (they encode vehicle geometry). Kits generate them offline with:

scripts/generate_crop_box_params.py <vehicle_info.param.yaml> <mirror.param.yaml> <output_dir>

and pass the outputs via the crop_box_filter_{self,secondary}_param_path arguments.

Scripts

  • septentrio_heading_converter.py — runtime node converting the septentrio AttEuler heading to a GnssInsOrientationStamped (installed).
  • generate_crop_box_params.py — offline generator for kit crop box files (not installed).

Package Dependencies

Please see <exec_depend> in package.xml.

CHANGELOG

Changelog for package autoware_sensing_launch

0.53.0 (2026-10-13)

  • Merge remote-tracking branch 'origin/main' into tmp/bot/bump_version_base

  • feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled (#1987)

    * feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled autoware_universe#13074 moves BlockageDiagComponent to agnocast_wrapper::Node, which is not designed to run inside a component container. Start the node as a standalone process when ENABLE_AGNOCAST=1 and keep loading it into target_container otherwise, as component_state_monitor.launch.py does for topic_state_monitor. The node name, namespace, remaps and parameters are identical in both branches.

    * refactor(autoware_sensing_launch): launch blockage_diag through the autoware_node action ---------

  • feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled (#1990)

    * feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled autoware_universe#13384 moves CropBoxFilterComponent to agnocast_wrapper::Node on an agnocast-only executor, which the component containers do not set up. Start each of the eight crop_box_filter nodes as a standalone process when ENABLE_AGNOCAST=1 and keep loading them into their container otherwise. The node names, namespaces, remaps and parameters are the same in both branches.

    * refactor(autoware_launch): launch crop_box_filter through the autoware_node action ---------

  • refactor(sensing)!: separate sensor driver and pre-processing layers from sensor kit configuration (#1938)

    * refactor(launch)!: rename tier4_sensing_launch to autoware_sensing_launch tier4_sensing_launch held one 19-line pass-through page that pushed the sensing namespace and included the sensor kit entry. The component page now includes that entry directly and pushes the namespace itself, so the layer is gone and the package is left as the home of the shared sensing units added next.

    • move the package to autoware_universe_launch/autoware_sensing_launch

    - rename tier4_sensing_component.launch.xml to components/component_sensing.launch.xml

    - drop the pass-through page and its diagram Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensing_launch): add the shared sensing unit library Declarative XML units for the sensor data pre/post-processing stages, each addressed by an absolute container path so a caller can place it in any container it owns.

    • pointcloud_container, the only unit that creates a container

    - pointcloud_crop_boxes, pointcloud_preprocessor_{cpu,cuda}, pointcloud_concatenation_{cpu,cuda}

    - blockage_diagnostics, gnss_postprocessor, imu_corrector, radar_tracks_postprocessor

    • shared preprocessing, blockage and radar defaults under config/

    - generate_crop_box_params.py, which turns vehicle info and mirror geometry into the crop box files the kits commit and pass in; septentrio_heading_converter.py is installed as a runtime node The parameter files and the heading converter are ported from tier4/aip_launcher (aip_common_sensor_launch), Apache-2.0 under the same TIER IV copyright. Nothing includes these units yet. Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensor_driver_launch): add the sensor driver layer The hardware interface layer, kept apart from the pre-processing units so that a sensor make and model can be brought into the ROS graph on its own.

    - nebula_{velodyne,robosense,hesai}: per-make driver chains, configured by the model param file shipped in the nebula driver package

    - velodyne_{VLP16,VLS128}, robosense_{Bpearl,Helios}, hesai_{XT32,OT128}: per-model wrappers

    - every file loads its node into target_container, an existing container owned by the caller; this package never creates one The

File truncated at 100 lines see the full file

Launch files

  • launch/blockage_diagnostics.launch.xml
      • target_container
      • node_name [default: blockage_diag]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • output/pointcloud [default: blockage_diag/pointcloud]
      • cloud_min_angle [default: 0.0]
      • cloud_max_angle [default: 360.0]
      • angle_range [default: [$(var cloud_min_angle), $(var cloud_max_angle)]]
      • max_range [default: 300.0]
      • horizontal_ring_id [default: 64]
      • vertical_bins [default: 128]
      • is_channel_order_top2down [default: true]
      • horizontal_resolution [default: 0.4]
      • blockage_diagnostics_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/blockage_diagnostics.param.yaml]
  • launch/gnss_postprocessor.launch.xml
      • input/fix [default: septentrio/nav_sat_fix]
      • input/atteuler [default: septentrio/atteuler]
      • input/orientation [default: septentrio/orientation]
      • output/pose [default: pose]
      • output/pose_with_covariance [default: pose_with_covariance]
      • output/fixed [default: fixed]
  • launch/imu_corrector.launch.xml
      • input/imu_raw
      • output/imu_data [default: imu_data]
      • input/odometry [default: /localization/kinematic_state]
      • input/pose_with_covariance [default: /localization/pose_estimator/pose_with_covariance]
      • output/imu_scaled [default: /sensing/imu_output_scaled]
      • imu_corrector_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/imu_corrector.param.yaml]
      • gyro_scale_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_scale_estimator.param.yaml]
      • gyro_bias_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_bias_estimator.param.yaml]
  • launch/pointcloud_concatenation_cpu.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • use_intra_process [default: false]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_concatenation_cuda.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_container.launch.xml
      • container_namespace
      • container_name [default: pointcloud_container]
      • use_multithread [default: false]
  • launch/pointcloud_crop_boxes.launch.xml
      • target_container
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • intermediate/pointcloud [default: self_cropped/pointcloud_ex]
      • output/pointcloud [default: $(var secondary_crop_box)_cropped/pointcloud_ex]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
  • launch/pointcloud_preprocessor_cpu.launch.xml
      • target_container
      • frame_id [default: lidar]
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • output_as_sensor_frame [default: true]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/pointcloud_preprocessor_cuda.launch.xml
      • target_container
      • enable_ring_outlier_filter [default: true]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/radar_tracks_postprocessor.launch.xml
      • input/tracks [default: objects_raw]
      • input/odometry [default: /localization/kinematic_state]
      • output/noise_tracks [default: noise_objects]
      • output/filtered_tracks [default: filtered_objects]
      • output/radar_detected_objects [default: detected_objects]
      • output/radar_tracked_objects [default: tracked_objects]
      • radar_tracks_msgs_converter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_msgs_converter.param.yaml]
      • radar_tracks_noise_filter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_noise_filter.param.yaml]

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged autoware_sensing_launch at Robotics Stack Exchange

Package symbol

autoware_sensing_launch package from autoware_launch repo

autoware_launch autoware_sample_designs autoware_control_config autoware_control_launch autoware_localization_config autoware_localization_launch autoware_map_config autoware_map_launch autoware_perception_config autoware_perception_launch autoware_planning_config autoware_planning_launch autoware_sensing_launch autoware_sensor_driver_launch awsim_labs_sensor_kit_description awsim_labs_sensor_kit_launch awsim_sensor_kit_description awsim_sensor_kit_launch carla_sensor_kit_description carla_sensor_kit_launch sample_sensor_kit_description sample_sensor_kit_launch single_lidar_sensor_kit_description single_lidar_sensor_kit_launch tier4_autoware_api_launch tier4_simulator_launch tier4_system_launch tier4_vehicle_launch awsim_labs_vehicle_description awsim_labs_vehicle_launch sample_vehicle_description sample_vehicle_launch

ROS Distro
github

API Docs
Browse Code

Package Summary

Version 0.53.0
License Apache License 2.0
Build type AMENT_CMAKE
Use RECOMMENDED

Repository Summary

Description
Checkout URI https://github.com/autowarefoundation/autoware_launch.git
VCS Type git
VCS Version main
Last Updated 2026-10-26
Dev Status UNKNOWN
Released UNRELEASED
Contributing Help Wanted (-)
Good First Issues (-)
Pull Requests to Review (-)

Package Description

Sensor data pre-processing pipelines shared by sensor kits

Maintainers

  • David Wong
  • Max Schmeller
  • Dai Nguyen

Authors

No additional authors.
autoware_universe_launch/autoware_sensing_launch/README.md

autoware_sensing_launch

Shared sensing launch library. Sensor kits own the sensing entry point (<sensor_model>_launch/launch/sensing.launch.xml, included by component_sensing.launch.xml); this package hosts the reusable pre/post-processing units the kits build on. Driver launch files live in the driver layer package autoware_sensor_driver_launch.

Container ownership and addressing

Sensor kits create the containers; units only load nodes into them. pointcloud_container.launch.xml is the one unit that creates a container, and it loads nothing into it.

Units that load composable nodes take the target container as an absolute path (target_container) supplied by the caller; units never derive it from their own namespace. Callers construct the path from their literal namespace knowledge and must keep it in step with their namespace pushes. The one exception is the concatenation units, whose pointcloud_container_name keeps its historical name and receives the top-level pointcloud container path passed through unchanged.

Launch files

  • pointcloud_container.launch.xml — component container for one sensor unit, created at container_namespace/container_name; use_multithread selects component_container_mt.
  • pointcloud_crop_boxes.launch.xml — vehicle body crop box followed by a secondary crop box, loaded into an existing container. Used on its own where the remaining preprocessing stages do not apply.
  • pointcloud_preprocessor_cpu.launch.xml — per-lidar preprocessing chain: the crop box unit, then distortion correction and ring outlier filter, loaded into an existing container. The secondary crop box is named by secondary_crop_box (e.g. mirror, wheels).
  • pointcloud_preprocessor_cuda.launch.xml — the same chain folded into one CUDA node; crop box stages come as folded lists in crop_box_param_path.
  • pointcloud_concatenation_cpu.launch.xml, pointcloud_concatenation_cuda.launch.xml — multi-lidar concatenation with time synchronization; the kit passes its concatenate_and_time_sync_node_param_path (input topic list is kit-specific).
  • blockage_diagnostics.launch.xml — blockage diagnostics on the raw pointcloud; angle_range defaults to the configured device angles (float-formatted).
  • gnss_postprocessor.launch.xml — septentrio heading converter plus NavSatFix-to-MGRS pose (autoware_gnss_poser).
  • imu_corrector.launch.xml — imu corrector plus gyro bias estimator (autoware_imu_corrector).
  • radar_tracks_postprocessor.launch.xml — radar tracks noise filter plus message converter.

Config

  • distortion_corrector_node.param.yaml, ring_outlier_filter_node.param.yaml — shared preprocessing defaults; kits with different sensors or thresholds pass their own file via the *_param_path arguments (the outlier shape parameters such as vertical_bins only take effect when publish_outlier_pointcloud is enabled).
  • blockage_diagnostics.param.yaml — blockage diagnostics defaults; runtime shape parameters (angle range, bins) are set per sensor via launch arguments.
  • radar_tracks_msgs_converter.param.yaml, radar_tracks_noise_filter.param.yaml — radar postprocessing defaults.

Crop box param files are owned by the sensor kits (they encode vehicle geometry). Kits generate them offline with:

scripts/generate_crop_box_params.py <vehicle_info.param.yaml> <mirror.param.yaml> <output_dir>

and pass the outputs via the crop_box_filter_{self,secondary}_param_path arguments.

Scripts

  • septentrio_heading_converter.py — runtime node converting the septentrio AttEuler heading to a GnssInsOrientationStamped (installed).
  • generate_crop_box_params.py — offline generator for kit crop box files (not installed).

Package Dependencies

Please see <exec_depend> in package.xml.

CHANGELOG

Changelog for package autoware_sensing_launch

0.53.0 (2026-10-13)

  • Merge remote-tracking branch 'origin/main' into tmp/bot/bump_version_base

  • feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled (#1987)

    * feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled autoware_universe#13074 moves BlockageDiagComponent to agnocast_wrapper::Node, which is not designed to run inside a component container. Start the node as a standalone process when ENABLE_AGNOCAST=1 and keep loading it into target_container otherwise, as component_state_monitor.launch.py does for topic_state_monitor. The node name, namespace, remaps and parameters are identical in both branches.

    * refactor(autoware_sensing_launch): launch blockage_diag through the autoware_node action ---------

  • feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled (#1990)

    * feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled autoware_universe#13384 moves CropBoxFilterComponent to agnocast_wrapper::Node on an agnocast-only executor, which the component containers do not set up. Start each of the eight crop_box_filter nodes as a standalone process when ENABLE_AGNOCAST=1 and keep loading them into their container otherwise. The node names, namespaces, remaps and parameters are the same in both branches.

    * refactor(autoware_launch): launch crop_box_filter through the autoware_node action ---------

  • refactor(sensing)!: separate sensor driver and pre-processing layers from sensor kit configuration (#1938)

    * refactor(launch)!: rename tier4_sensing_launch to autoware_sensing_launch tier4_sensing_launch held one 19-line pass-through page that pushed the sensing namespace and included the sensor kit entry. The component page now includes that entry directly and pushes the namespace itself, so the layer is gone and the package is left as the home of the shared sensing units added next.

    • move the package to autoware_universe_launch/autoware_sensing_launch

    - rename tier4_sensing_component.launch.xml to components/component_sensing.launch.xml

    - drop the pass-through page and its diagram Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensing_launch): add the shared sensing unit library Declarative XML units for the sensor data pre/post-processing stages, each addressed by an absolute container path so a caller can place it in any container it owns.

    • pointcloud_container, the only unit that creates a container

    - pointcloud_crop_boxes, pointcloud_preprocessor_{cpu,cuda}, pointcloud_concatenation_{cpu,cuda}

    - blockage_diagnostics, gnss_postprocessor, imu_corrector, radar_tracks_postprocessor

    • shared preprocessing, blockage and radar defaults under config/

    - generate_crop_box_params.py, which turns vehicle info and mirror geometry into the crop box files the kits commit and pass in; septentrio_heading_converter.py is installed as a runtime node The parameter files and the heading converter are ported from tier4/aip_launcher (aip_common_sensor_launch), Apache-2.0 under the same TIER IV copyright. Nothing includes these units yet. Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensor_driver_launch): add the sensor driver layer The hardware interface layer, kept apart from the pre-processing units so that a sensor make and model can be brought into the ROS graph on its own.

    - nebula_{velodyne,robosense,hesai}: per-make driver chains, configured by the model param file shipped in the nebula driver package

    - velodyne_{VLP16,VLS128}, robosense_{Bpearl,Helios}, hesai_{XT32,OT128}: per-model wrappers

    - every file loads its node into target_container, an existing container owned by the caller; this package never creates one The

File truncated at 100 lines see the full file

Launch files

  • launch/blockage_diagnostics.launch.xml
      • target_container
      • node_name [default: blockage_diag]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • output/pointcloud [default: blockage_diag/pointcloud]
      • cloud_min_angle [default: 0.0]
      • cloud_max_angle [default: 360.0]
      • angle_range [default: [$(var cloud_min_angle), $(var cloud_max_angle)]]
      • max_range [default: 300.0]
      • horizontal_ring_id [default: 64]
      • vertical_bins [default: 128]
      • is_channel_order_top2down [default: true]
      • horizontal_resolution [default: 0.4]
      • blockage_diagnostics_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/blockage_diagnostics.param.yaml]
  • launch/gnss_postprocessor.launch.xml
      • input/fix [default: septentrio/nav_sat_fix]
      • input/atteuler [default: septentrio/atteuler]
      • input/orientation [default: septentrio/orientation]
      • output/pose [default: pose]
      • output/pose_with_covariance [default: pose_with_covariance]
      • output/fixed [default: fixed]
  • launch/imu_corrector.launch.xml
      • input/imu_raw
      • output/imu_data [default: imu_data]
      • input/odometry [default: /localization/kinematic_state]
      • input/pose_with_covariance [default: /localization/pose_estimator/pose_with_covariance]
      • output/imu_scaled [default: /sensing/imu_output_scaled]
      • imu_corrector_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/imu_corrector.param.yaml]
      • gyro_scale_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_scale_estimator.param.yaml]
      • gyro_bias_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_bias_estimator.param.yaml]
  • launch/pointcloud_concatenation_cpu.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • use_intra_process [default: false]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_concatenation_cuda.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_container.launch.xml
      • container_namespace
      • container_name [default: pointcloud_container]
      • use_multithread [default: false]
  • launch/pointcloud_crop_boxes.launch.xml
      • target_container
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • intermediate/pointcloud [default: self_cropped/pointcloud_ex]
      • output/pointcloud [default: $(var secondary_crop_box)_cropped/pointcloud_ex]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
  • launch/pointcloud_preprocessor_cpu.launch.xml
      • target_container
      • frame_id [default: lidar]
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • output_as_sensor_frame [default: true]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/pointcloud_preprocessor_cuda.launch.xml
      • target_container
      • enable_ring_outlier_filter [default: true]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/radar_tracks_postprocessor.launch.xml
      • input/tracks [default: objects_raw]
      • input/odometry [default: /localization/kinematic_state]
      • output/noise_tracks [default: noise_objects]
      • output/filtered_tracks [default: filtered_objects]
      • output/radar_detected_objects [default: detected_objects]
      • output/radar_tracked_objects [default: tracked_objects]
      • radar_tracks_msgs_converter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_msgs_converter.param.yaml]
      • radar_tracks_noise_filter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_noise_filter.param.yaml]

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged autoware_sensing_launch at Robotics Stack Exchange

No version for distro galactic showing github. Known supported distros are highlighted in the buttons above.
Package symbol

autoware_sensing_launch package from autoware_launch repo

autoware_launch autoware_sample_designs autoware_control_config autoware_control_launch autoware_localization_config autoware_localization_launch autoware_map_config autoware_map_launch autoware_perception_config autoware_perception_launch autoware_planning_config autoware_planning_launch autoware_sensing_launch autoware_sensor_driver_launch awsim_labs_sensor_kit_description awsim_labs_sensor_kit_launch awsim_sensor_kit_description awsim_sensor_kit_launch carla_sensor_kit_description carla_sensor_kit_launch sample_sensor_kit_description sample_sensor_kit_launch single_lidar_sensor_kit_description single_lidar_sensor_kit_launch tier4_autoware_api_launch tier4_simulator_launch tier4_system_launch tier4_vehicle_launch awsim_labs_vehicle_description awsim_labs_vehicle_launch sample_vehicle_description sample_vehicle_launch

ROS Distro
github

API Docs
Browse Code

Package Summary

Version 0.53.0
License Apache License 2.0
Build type AMENT_CMAKE
Use RECOMMENDED

Repository Summary

Description
Checkout URI https://github.com/autowarefoundation/autoware_launch.git
VCS Type git
VCS Version main
Last Updated 2026-10-26
Dev Status UNKNOWN
Released UNRELEASED
Contributing Help Wanted (-)
Good First Issues (-)
Pull Requests to Review (-)

Package Description

Sensor data pre-processing pipelines shared by sensor kits

Maintainers

  • David Wong
  • Max Schmeller
  • Dai Nguyen

Authors

No additional authors.
autoware_universe_launch/autoware_sensing_launch/README.md

autoware_sensing_launch

Shared sensing launch library. Sensor kits own the sensing entry point (<sensor_model>_launch/launch/sensing.launch.xml, included by component_sensing.launch.xml); this package hosts the reusable pre/post-processing units the kits build on. Driver launch files live in the driver layer package autoware_sensor_driver_launch.

Container ownership and addressing

Sensor kits create the containers; units only load nodes into them. pointcloud_container.launch.xml is the one unit that creates a container, and it loads nothing into it.

Units that load composable nodes take the target container as an absolute path (target_container) supplied by the caller; units never derive it from their own namespace. Callers construct the path from their literal namespace knowledge and must keep it in step with their namespace pushes. The one exception is the concatenation units, whose pointcloud_container_name keeps its historical name and receives the top-level pointcloud container path passed through unchanged.

Launch files

  • pointcloud_container.launch.xml — component container for one sensor unit, created at container_namespace/container_name; use_multithread selects component_container_mt.
  • pointcloud_crop_boxes.launch.xml — vehicle body crop box followed by a secondary crop box, loaded into an existing container. Used on its own where the remaining preprocessing stages do not apply.
  • pointcloud_preprocessor_cpu.launch.xml — per-lidar preprocessing chain: the crop box unit, then distortion correction and ring outlier filter, loaded into an existing container. The secondary crop box is named by secondary_crop_box (e.g. mirror, wheels).
  • pointcloud_preprocessor_cuda.launch.xml — the same chain folded into one CUDA node; crop box stages come as folded lists in crop_box_param_path.
  • pointcloud_concatenation_cpu.launch.xml, pointcloud_concatenation_cuda.launch.xml — multi-lidar concatenation with time synchronization; the kit passes its concatenate_and_time_sync_node_param_path (input topic list is kit-specific).
  • blockage_diagnostics.launch.xml — blockage diagnostics on the raw pointcloud; angle_range defaults to the configured device angles (float-formatted).
  • gnss_postprocessor.launch.xml — septentrio heading converter plus NavSatFix-to-MGRS pose (autoware_gnss_poser).
  • imu_corrector.launch.xml — imu corrector plus gyro bias estimator (autoware_imu_corrector).
  • radar_tracks_postprocessor.launch.xml — radar tracks noise filter plus message converter.

Config

  • distortion_corrector_node.param.yaml, ring_outlier_filter_node.param.yaml — shared preprocessing defaults; kits with different sensors or thresholds pass their own file via the *_param_path arguments (the outlier shape parameters such as vertical_bins only take effect when publish_outlier_pointcloud is enabled).
  • blockage_diagnostics.param.yaml — blockage diagnostics defaults; runtime shape parameters (angle range, bins) are set per sensor via launch arguments.
  • radar_tracks_msgs_converter.param.yaml, radar_tracks_noise_filter.param.yaml — radar postprocessing defaults.

Crop box param files are owned by the sensor kits (they encode vehicle geometry). Kits generate them offline with:

scripts/generate_crop_box_params.py <vehicle_info.param.yaml> <mirror.param.yaml> <output_dir>

and pass the outputs via the crop_box_filter_{self,secondary}_param_path arguments.

Scripts

  • septentrio_heading_converter.py — runtime node converting the septentrio AttEuler heading to a GnssInsOrientationStamped (installed).
  • generate_crop_box_params.py — offline generator for kit crop box files (not installed).

Package Dependencies

Please see <exec_depend> in package.xml.

CHANGELOG

Changelog for package autoware_sensing_launch

0.53.0 (2026-10-13)

  • Merge remote-tracking branch 'origin/main' into tmp/bot/bump_version_base

  • feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled (#1987)

    * feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled autoware_universe#13074 moves BlockageDiagComponent to agnocast_wrapper::Node, which is not designed to run inside a component container. Start the node as a standalone process when ENABLE_AGNOCAST=1 and keep loading it into target_container otherwise, as component_state_monitor.launch.py does for topic_state_monitor. The node name, namespace, remaps and parameters are identical in both branches.

    * refactor(autoware_sensing_launch): launch blockage_diag through the autoware_node action ---------

  • feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled (#1990)

    * feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled autoware_universe#13384 moves CropBoxFilterComponent to agnocast_wrapper::Node on an agnocast-only executor, which the component containers do not set up. Start each of the eight crop_box_filter nodes as a standalone process when ENABLE_AGNOCAST=1 and keep loading them into their container otherwise. The node names, namespaces, remaps and parameters are the same in both branches.

    * refactor(autoware_launch): launch crop_box_filter through the autoware_node action ---------

  • refactor(sensing)!: separate sensor driver and pre-processing layers from sensor kit configuration (#1938)

    * refactor(launch)!: rename tier4_sensing_launch to autoware_sensing_launch tier4_sensing_launch held one 19-line pass-through page that pushed the sensing namespace and included the sensor kit entry. The component page now includes that entry directly and pushes the namespace itself, so the layer is gone and the package is left as the home of the shared sensing units added next.

    • move the package to autoware_universe_launch/autoware_sensing_launch

    - rename tier4_sensing_component.launch.xml to components/component_sensing.launch.xml

    - drop the pass-through page and its diagram Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensing_launch): add the shared sensing unit library Declarative XML units for the sensor data pre/post-processing stages, each addressed by an absolute container path so a caller can place it in any container it owns.

    • pointcloud_container, the only unit that creates a container

    - pointcloud_crop_boxes, pointcloud_preprocessor_{cpu,cuda}, pointcloud_concatenation_{cpu,cuda}

    - blockage_diagnostics, gnss_postprocessor, imu_corrector, radar_tracks_postprocessor

    • shared preprocessing, blockage and radar defaults under config/

    - generate_crop_box_params.py, which turns vehicle info and mirror geometry into the crop box files the kits commit and pass in; septentrio_heading_converter.py is installed as a runtime node The parameter files and the heading converter are ported from tier4/aip_launcher (aip_common_sensor_launch), Apache-2.0 under the same TIER IV copyright. Nothing includes these units yet. Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensor_driver_launch): add the sensor driver layer The hardware interface layer, kept apart from the pre-processing units so that a sensor make and model can be brought into the ROS graph on its own.

    - nebula_{velodyne,robosense,hesai}: per-make driver chains, configured by the model param file shipped in the nebula driver package

    - velodyne_{VLP16,VLS128}, robosense_{Bpearl,Helios}, hesai_{XT32,OT128}: per-model wrappers

    - every file loads its node into target_container, an existing container owned by the caller; this package never creates one The

File truncated at 100 lines see the full file

Launch files

  • launch/blockage_diagnostics.launch.xml
      • target_container
      • node_name [default: blockage_diag]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • output/pointcloud [default: blockage_diag/pointcloud]
      • cloud_min_angle [default: 0.0]
      • cloud_max_angle [default: 360.0]
      • angle_range [default: [$(var cloud_min_angle), $(var cloud_max_angle)]]
      • max_range [default: 300.0]
      • horizontal_ring_id [default: 64]
      • vertical_bins [default: 128]
      • is_channel_order_top2down [default: true]
      • horizontal_resolution [default: 0.4]
      • blockage_diagnostics_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/blockage_diagnostics.param.yaml]
  • launch/gnss_postprocessor.launch.xml
      • input/fix [default: septentrio/nav_sat_fix]
      • input/atteuler [default: septentrio/atteuler]
      • input/orientation [default: septentrio/orientation]
      • output/pose [default: pose]
      • output/pose_with_covariance [default: pose_with_covariance]
      • output/fixed [default: fixed]
  • launch/imu_corrector.launch.xml
      • input/imu_raw
      • output/imu_data [default: imu_data]
      • input/odometry [default: /localization/kinematic_state]
      • input/pose_with_covariance [default: /localization/pose_estimator/pose_with_covariance]
      • output/imu_scaled [default: /sensing/imu_output_scaled]
      • imu_corrector_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/imu_corrector.param.yaml]
      • gyro_scale_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_scale_estimator.param.yaml]
      • gyro_bias_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_bias_estimator.param.yaml]
  • launch/pointcloud_concatenation_cpu.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • use_intra_process [default: false]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_concatenation_cuda.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_container.launch.xml
      • container_namespace
      • container_name [default: pointcloud_container]
      • use_multithread [default: false]
  • launch/pointcloud_crop_boxes.launch.xml
      • target_container
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • intermediate/pointcloud [default: self_cropped/pointcloud_ex]
      • output/pointcloud [default: $(var secondary_crop_box)_cropped/pointcloud_ex]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
  • launch/pointcloud_preprocessor_cpu.launch.xml
      • target_container
      • frame_id [default: lidar]
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • output_as_sensor_frame [default: true]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/pointcloud_preprocessor_cuda.launch.xml
      • target_container
      • enable_ring_outlier_filter [default: true]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/radar_tracks_postprocessor.launch.xml
      • input/tracks [default: objects_raw]
      • input/odometry [default: /localization/kinematic_state]
      • output/noise_tracks [default: noise_objects]
      • output/filtered_tracks [default: filtered_objects]
      • output/radar_detected_objects [default: detected_objects]
      • output/radar_tracked_objects [default: tracked_objects]
      • radar_tracks_msgs_converter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_msgs_converter.param.yaml]
      • radar_tracks_noise_filter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_noise_filter.param.yaml]

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged autoware_sensing_launch at Robotics Stack Exchange

No version for distro iron showing github. Known supported distros are highlighted in the buttons above.
Package symbol

autoware_sensing_launch package from autoware_launch repo

autoware_launch autoware_sample_designs autoware_control_config autoware_control_launch autoware_localization_config autoware_localization_launch autoware_map_config autoware_map_launch autoware_perception_config autoware_perception_launch autoware_planning_config autoware_planning_launch autoware_sensing_launch autoware_sensor_driver_launch awsim_labs_sensor_kit_description awsim_labs_sensor_kit_launch awsim_sensor_kit_description awsim_sensor_kit_launch carla_sensor_kit_description carla_sensor_kit_launch sample_sensor_kit_description sample_sensor_kit_launch single_lidar_sensor_kit_description single_lidar_sensor_kit_launch tier4_autoware_api_launch tier4_simulator_launch tier4_system_launch tier4_vehicle_launch awsim_labs_vehicle_description awsim_labs_vehicle_launch sample_vehicle_description sample_vehicle_launch

ROS Distro
github

API Docs
Browse Code

Package Summary

Version 0.53.0
License Apache License 2.0
Build type AMENT_CMAKE
Use RECOMMENDED

Repository Summary

Description
Checkout URI https://github.com/autowarefoundation/autoware_launch.git
VCS Type git
VCS Version main
Last Updated 2026-10-26
Dev Status UNKNOWN
Released UNRELEASED
Contributing Help Wanted (-)
Good First Issues (-)
Pull Requests to Review (-)

Package Description

Sensor data pre-processing pipelines shared by sensor kits

Maintainers

  • David Wong
  • Max Schmeller
  • Dai Nguyen

Authors

No additional authors.
autoware_universe_launch/autoware_sensing_launch/README.md

autoware_sensing_launch

Shared sensing launch library. Sensor kits own the sensing entry point (<sensor_model>_launch/launch/sensing.launch.xml, included by component_sensing.launch.xml); this package hosts the reusable pre/post-processing units the kits build on. Driver launch files live in the driver layer package autoware_sensor_driver_launch.

Container ownership and addressing

Sensor kits create the containers; units only load nodes into them. pointcloud_container.launch.xml is the one unit that creates a container, and it loads nothing into it.

Units that load composable nodes take the target container as an absolute path (target_container) supplied by the caller; units never derive it from their own namespace. Callers construct the path from their literal namespace knowledge and must keep it in step with their namespace pushes. The one exception is the concatenation units, whose pointcloud_container_name keeps its historical name and receives the top-level pointcloud container path passed through unchanged.

Launch files

  • pointcloud_container.launch.xml — component container for one sensor unit, created at container_namespace/container_name; use_multithread selects component_container_mt.
  • pointcloud_crop_boxes.launch.xml — vehicle body crop box followed by a secondary crop box, loaded into an existing container. Used on its own where the remaining preprocessing stages do not apply.
  • pointcloud_preprocessor_cpu.launch.xml — per-lidar preprocessing chain: the crop box unit, then distortion correction and ring outlier filter, loaded into an existing container. The secondary crop box is named by secondary_crop_box (e.g. mirror, wheels).
  • pointcloud_preprocessor_cuda.launch.xml — the same chain folded into one CUDA node; crop box stages come as folded lists in crop_box_param_path.
  • pointcloud_concatenation_cpu.launch.xml, pointcloud_concatenation_cuda.launch.xml — multi-lidar concatenation with time synchronization; the kit passes its concatenate_and_time_sync_node_param_path (input topic list is kit-specific).
  • blockage_diagnostics.launch.xml — blockage diagnostics on the raw pointcloud; angle_range defaults to the configured device angles (float-formatted).
  • gnss_postprocessor.launch.xml — septentrio heading converter plus NavSatFix-to-MGRS pose (autoware_gnss_poser).
  • imu_corrector.launch.xml — imu corrector plus gyro bias estimator (autoware_imu_corrector).
  • radar_tracks_postprocessor.launch.xml — radar tracks noise filter plus message converter.

Config

  • distortion_corrector_node.param.yaml, ring_outlier_filter_node.param.yaml — shared preprocessing defaults; kits with different sensors or thresholds pass their own file via the *_param_path arguments (the outlier shape parameters such as vertical_bins only take effect when publish_outlier_pointcloud is enabled).
  • blockage_diagnostics.param.yaml — blockage diagnostics defaults; runtime shape parameters (angle range, bins) are set per sensor via launch arguments.
  • radar_tracks_msgs_converter.param.yaml, radar_tracks_noise_filter.param.yaml — radar postprocessing defaults.

Crop box param files are owned by the sensor kits (they encode vehicle geometry). Kits generate them offline with:

scripts/generate_crop_box_params.py <vehicle_info.param.yaml> <mirror.param.yaml> <output_dir>

and pass the outputs via the crop_box_filter_{self,secondary}_param_path arguments.

Scripts

  • septentrio_heading_converter.py — runtime node converting the septentrio AttEuler heading to a GnssInsOrientationStamped (installed).
  • generate_crop_box_params.py — offline generator for kit crop box files (not installed).

Package Dependencies

Please see <exec_depend> in package.xml.

CHANGELOG

Changelog for package autoware_sensing_launch

0.53.0 (2026-10-13)

  • Merge remote-tracking branch 'origin/main' into tmp/bot/bump_version_base

  • feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled (#1987)

    * feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled autoware_universe#13074 moves BlockageDiagComponent to agnocast_wrapper::Node, which is not designed to run inside a component container. Start the node as a standalone process when ENABLE_AGNOCAST=1 and keep loading it into target_container otherwise, as component_state_monitor.launch.py does for topic_state_monitor. The node name, namespace, remaps and parameters are identical in both branches.

    * refactor(autoware_sensing_launch): launch blockage_diag through the autoware_node action ---------

  • feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled (#1990)

    * feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled autoware_universe#13384 moves CropBoxFilterComponent to agnocast_wrapper::Node on an agnocast-only executor, which the component containers do not set up. Start each of the eight crop_box_filter nodes as a standalone process when ENABLE_AGNOCAST=1 and keep loading them into their container otherwise. The node names, namespaces, remaps and parameters are the same in both branches.

    * refactor(autoware_launch): launch crop_box_filter through the autoware_node action ---------

  • refactor(sensing)!: separate sensor driver and pre-processing layers from sensor kit configuration (#1938)

    * refactor(launch)!: rename tier4_sensing_launch to autoware_sensing_launch tier4_sensing_launch held one 19-line pass-through page that pushed the sensing namespace and included the sensor kit entry. The component page now includes that entry directly and pushes the namespace itself, so the layer is gone and the package is left as the home of the shared sensing units added next.

    • move the package to autoware_universe_launch/autoware_sensing_launch

    - rename tier4_sensing_component.launch.xml to components/component_sensing.launch.xml

    - drop the pass-through page and its diagram Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensing_launch): add the shared sensing unit library Declarative XML units for the sensor data pre/post-processing stages, each addressed by an absolute container path so a caller can place it in any container it owns.

    • pointcloud_container, the only unit that creates a container

    - pointcloud_crop_boxes, pointcloud_preprocessor_{cpu,cuda}, pointcloud_concatenation_{cpu,cuda}

    - blockage_diagnostics, gnss_postprocessor, imu_corrector, radar_tracks_postprocessor

    • shared preprocessing, blockage and radar defaults under config/

    - generate_crop_box_params.py, which turns vehicle info and mirror geometry into the crop box files the kits commit and pass in; septentrio_heading_converter.py is installed as a runtime node The parameter files and the heading converter are ported from tier4/aip_launcher (aip_common_sensor_launch), Apache-2.0 under the same TIER IV copyright. Nothing includes these units yet. Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensor_driver_launch): add the sensor driver layer The hardware interface layer, kept apart from the pre-processing units so that a sensor make and model can be brought into the ROS graph on its own.

    - nebula_{velodyne,robosense,hesai}: per-make driver chains, configured by the model param file shipped in the nebula driver package

    - velodyne_{VLP16,VLS128}, robosense_{Bpearl,Helios}, hesai_{XT32,OT128}: per-model wrappers

    - every file loads its node into target_container, an existing container owned by the caller; this package never creates one The

File truncated at 100 lines see the full file

Launch files

  • launch/blockage_diagnostics.launch.xml
      • target_container
      • node_name [default: blockage_diag]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • output/pointcloud [default: blockage_diag/pointcloud]
      • cloud_min_angle [default: 0.0]
      • cloud_max_angle [default: 360.0]
      • angle_range [default: [$(var cloud_min_angle), $(var cloud_max_angle)]]
      • max_range [default: 300.0]
      • horizontal_ring_id [default: 64]
      • vertical_bins [default: 128]
      • is_channel_order_top2down [default: true]
      • horizontal_resolution [default: 0.4]
      • blockage_diagnostics_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/blockage_diagnostics.param.yaml]
  • launch/gnss_postprocessor.launch.xml
      • input/fix [default: septentrio/nav_sat_fix]
      • input/atteuler [default: septentrio/atteuler]
      • input/orientation [default: septentrio/orientation]
      • output/pose [default: pose]
      • output/pose_with_covariance [default: pose_with_covariance]
      • output/fixed [default: fixed]
  • launch/imu_corrector.launch.xml
      • input/imu_raw
      • output/imu_data [default: imu_data]
      • input/odometry [default: /localization/kinematic_state]
      • input/pose_with_covariance [default: /localization/pose_estimator/pose_with_covariance]
      • output/imu_scaled [default: /sensing/imu_output_scaled]
      • imu_corrector_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/imu_corrector.param.yaml]
      • gyro_scale_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_scale_estimator.param.yaml]
      • gyro_bias_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_bias_estimator.param.yaml]
  • launch/pointcloud_concatenation_cpu.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • use_intra_process [default: false]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_concatenation_cuda.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_container.launch.xml
      • container_namespace
      • container_name [default: pointcloud_container]
      • use_multithread [default: false]
  • launch/pointcloud_crop_boxes.launch.xml
      • target_container
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • intermediate/pointcloud [default: self_cropped/pointcloud_ex]
      • output/pointcloud [default: $(var secondary_crop_box)_cropped/pointcloud_ex]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
  • launch/pointcloud_preprocessor_cpu.launch.xml
      • target_container
      • frame_id [default: lidar]
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • output_as_sensor_frame [default: true]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/pointcloud_preprocessor_cuda.launch.xml
      • target_container
      • enable_ring_outlier_filter [default: true]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/radar_tracks_postprocessor.launch.xml
      • input/tracks [default: objects_raw]
      • input/odometry [default: /localization/kinematic_state]
      • output/noise_tracks [default: noise_objects]
      • output/filtered_tracks [default: filtered_objects]
      • output/radar_detected_objects [default: detected_objects]
      • output/radar_tracked_objects [default: tracked_objects]
      • radar_tracks_msgs_converter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_msgs_converter.param.yaml]
      • radar_tracks_noise_filter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_noise_filter.param.yaml]

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged autoware_sensing_launch at Robotics Stack Exchange

No version for distro melodic showing github. Known supported distros are highlighted in the buttons above.
Package symbol

autoware_sensing_launch package from autoware_launch repo

autoware_launch autoware_sample_designs autoware_control_config autoware_control_launch autoware_localization_config autoware_localization_launch autoware_map_config autoware_map_launch autoware_perception_config autoware_perception_launch autoware_planning_config autoware_planning_launch autoware_sensing_launch autoware_sensor_driver_launch awsim_labs_sensor_kit_description awsim_labs_sensor_kit_launch awsim_sensor_kit_description awsim_sensor_kit_launch carla_sensor_kit_description carla_sensor_kit_launch sample_sensor_kit_description sample_sensor_kit_launch single_lidar_sensor_kit_description single_lidar_sensor_kit_launch tier4_autoware_api_launch tier4_simulator_launch tier4_system_launch tier4_vehicle_launch awsim_labs_vehicle_description awsim_labs_vehicle_launch sample_vehicle_description sample_vehicle_launch

ROS Distro
github

API Docs
Browse Code
Wiki

Package Summary

Version 0.53.0
License Apache License 2.0
Build type AMENT_CMAKE
Use RECOMMENDED

Repository Summary

Description
Checkout URI https://github.com/autowarefoundation/autoware_launch.git
VCS Type git
VCS Version main
Last Updated 2026-10-26
Dev Status UNKNOWN
Released UNRELEASED
Contributing Help Wanted (-)
Good First Issues (-)
Pull Requests to Review (-)

Package Description

Sensor data pre-processing pipelines shared by sensor kits

Maintainers

  • David Wong
  • Max Schmeller
  • Dai Nguyen

Authors

No additional authors.
autoware_universe_launch/autoware_sensing_launch/README.md

autoware_sensing_launch

Shared sensing launch library. Sensor kits own the sensing entry point (<sensor_model>_launch/launch/sensing.launch.xml, included by component_sensing.launch.xml); this package hosts the reusable pre/post-processing units the kits build on. Driver launch files live in the driver layer package autoware_sensor_driver_launch.

Container ownership and addressing

Sensor kits create the containers; units only load nodes into them. pointcloud_container.launch.xml is the one unit that creates a container, and it loads nothing into it.

Units that load composable nodes take the target container as an absolute path (target_container) supplied by the caller; units never derive it from their own namespace. Callers construct the path from their literal namespace knowledge and must keep it in step with their namespace pushes. The one exception is the concatenation units, whose pointcloud_container_name keeps its historical name and receives the top-level pointcloud container path passed through unchanged.

Launch files

  • pointcloud_container.launch.xml — component container for one sensor unit, created at container_namespace/container_name; use_multithread selects component_container_mt.
  • pointcloud_crop_boxes.launch.xml — vehicle body crop box followed by a secondary crop box, loaded into an existing container. Used on its own where the remaining preprocessing stages do not apply.
  • pointcloud_preprocessor_cpu.launch.xml — per-lidar preprocessing chain: the crop box unit, then distortion correction and ring outlier filter, loaded into an existing container. The secondary crop box is named by secondary_crop_box (e.g. mirror, wheels).
  • pointcloud_preprocessor_cuda.launch.xml — the same chain folded into one CUDA node; crop box stages come as folded lists in crop_box_param_path.
  • pointcloud_concatenation_cpu.launch.xml, pointcloud_concatenation_cuda.launch.xml — multi-lidar concatenation with time synchronization; the kit passes its concatenate_and_time_sync_node_param_path (input topic list is kit-specific).
  • blockage_diagnostics.launch.xml — blockage diagnostics on the raw pointcloud; angle_range defaults to the configured device angles (float-formatted).
  • gnss_postprocessor.launch.xml — septentrio heading converter plus NavSatFix-to-MGRS pose (autoware_gnss_poser).
  • imu_corrector.launch.xml — imu corrector plus gyro bias estimator (autoware_imu_corrector).
  • radar_tracks_postprocessor.launch.xml — radar tracks noise filter plus message converter.

Config

  • distortion_corrector_node.param.yaml, ring_outlier_filter_node.param.yaml — shared preprocessing defaults; kits with different sensors or thresholds pass their own file via the *_param_path arguments (the outlier shape parameters such as vertical_bins only take effect when publish_outlier_pointcloud is enabled).
  • blockage_diagnostics.param.yaml — blockage diagnostics defaults; runtime shape parameters (angle range, bins) are set per sensor via launch arguments.
  • radar_tracks_msgs_converter.param.yaml, radar_tracks_noise_filter.param.yaml — radar postprocessing defaults.

Crop box param files are owned by the sensor kits (they encode vehicle geometry). Kits generate them offline with:

scripts/generate_crop_box_params.py <vehicle_info.param.yaml> <mirror.param.yaml> <output_dir>

and pass the outputs via the crop_box_filter_{self,secondary}_param_path arguments.

Scripts

  • septentrio_heading_converter.py — runtime node converting the septentrio AttEuler heading to a GnssInsOrientationStamped (installed).
  • generate_crop_box_params.py — offline generator for kit crop box files (not installed).

Package Dependencies

Please see <exec_depend> in package.xml.

CHANGELOG

Changelog for package autoware_sensing_launch

0.53.0 (2026-10-13)

  • Merge remote-tracking branch 'origin/main' into tmp/bot/bump_version_base

  • feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled (#1987)

    * feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled autoware_universe#13074 moves BlockageDiagComponent to agnocast_wrapper::Node, which is not designed to run inside a component container. Start the node as a standalone process when ENABLE_AGNOCAST=1 and keep loading it into target_container otherwise, as component_state_monitor.launch.py does for topic_state_monitor. The node name, namespace, remaps and parameters are identical in both branches.

    * refactor(autoware_sensing_launch): launch blockage_diag through the autoware_node action ---------

  • feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled (#1990)

    * feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled autoware_universe#13384 moves CropBoxFilterComponent to agnocast_wrapper::Node on an agnocast-only executor, which the component containers do not set up. Start each of the eight crop_box_filter nodes as a standalone process when ENABLE_AGNOCAST=1 and keep loading them into their container otherwise. The node names, namespaces, remaps and parameters are the same in both branches.

    * refactor(autoware_launch): launch crop_box_filter through the autoware_node action ---------

  • refactor(sensing)!: separate sensor driver and pre-processing layers from sensor kit configuration (#1938)

    * refactor(launch)!: rename tier4_sensing_launch to autoware_sensing_launch tier4_sensing_launch held one 19-line pass-through page that pushed the sensing namespace and included the sensor kit entry. The component page now includes that entry directly and pushes the namespace itself, so the layer is gone and the package is left as the home of the shared sensing units added next.

    • move the package to autoware_universe_launch/autoware_sensing_launch

    - rename tier4_sensing_component.launch.xml to components/component_sensing.launch.xml

    - drop the pass-through page and its diagram Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensing_launch): add the shared sensing unit library Declarative XML units for the sensor data pre/post-processing stages, each addressed by an absolute container path so a caller can place it in any container it owns.

    • pointcloud_container, the only unit that creates a container

    - pointcloud_crop_boxes, pointcloud_preprocessor_{cpu,cuda}, pointcloud_concatenation_{cpu,cuda}

    - blockage_diagnostics, gnss_postprocessor, imu_corrector, radar_tracks_postprocessor

    • shared preprocessing, blockage and radar defaults under config/

    - generate_crop_box_params.py, which turns vehicle info and mirror geometry into the crop box files the kits commit and pass in; septentrio_heading_converter.py is installed as a runtime node The parameter files and the heading converter are ported from tier4/aip_launcher (aip_common_sensor_launch), Apache-2.0 under the same TIER IV copyright. Nothing includes these units yet. Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensor_driver_launch): add the sensor driver layer The hardware interface layer, kept apart from the pre-processing units so that a sensor make and model can be brought into the ROS graph on its own.

    - nebula_{velodyne,robosense,hesai}: per-make driver chains, configured by the model param file shipped in the nebula driver package

    - velodyne_{VLP16,VLS128}, robosense_{Bpearl,Helios}, hesai_{XT32,OT128}: per-model wrappers

    - every file loads its node into target_container, an existing container owned by the caller; this package never creates one The

File truncated at 100 lines see the full file

Launch files

  • launch/blockage_diagnostics.launch.xml
      • target_container
      • node_name [default: blockage_diag]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • output/pointcloud [default: blockage_diag/pointcloud]
      • cloud_min_angle [default: 0.0]
      • cloud_max_angle [default: 360.0]
      • angle_range [default: [$(var cloud_min_angle), $(var cloud_max_angle)]]
      • max_range [default: 300.0]
      • horizontal_ring_id [default: 64]
      • vertical_bins [default: 128]
      • is_channel_order_top2down [default: true]
      • horizontal_resolution [default: 0.4]
      • blockage_diagnostics_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/blockage_diagnostics.param.yaml]
  • launch/gnss_postprocessor.launch.xml
      • input/fix [default: septentrio/nav_sat_fix]
      • input/atteuler [default: septentrio/atteuler]
      • input/orientation [default: septentrio/orientation]
      • output/pose [default: pose]
      • output/pose_with_covariance [default: pose_with_covariance]
      • output/fixed [default: fixed]
  • launch/imu_corrector.launch.xml
      • input/imu_raw
      • output/imu_data [default: imu_data]
      • input/odometry [default: /localization/kinematic_state]
      • input/pose_with_covariance [default: /localization/pose_estimator/pose_with_covariance]
      • output/imu_scaled [default: /sensing/imu_output_scaled]
      • imu_corrector_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/imu_corrector.param.yaml]
      • gyro_scale_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_scale_estimator.param.yaml]
      • gyro_bias_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_bias_estimator.param.yaml]
  • launch/pointcloud_concatenation_cpu.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • use_intra_process [default: false]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_concatenation_cuda.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_container.launch.xml
      • container_namespace
      • container_name [default: pointcloud_container]
      • use_multithread [default: false]
  • launch/pointcloud_crop_boxes.launch.xml
      • target_container
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • intermediate/pointcloud [default: self_cropped/pointcloud_ex]
      • output/pointcloud [default: $(var secondary_crop_box)_cropped/pointcloud_ex]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
  • launch/pointcloud_preprocessor_cpu.launch.xml
      • target_container
      • frame_id [default: lidar]
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • output_as_sensor_frame [default: true]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/pointcloud_preprocessor_cuda.launch.xml
      • target_container
      • enable_ring_outlier_filter [default: true]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/radar_tracks_postprocessor.launch.xml
      • input/tracks [default: objects_raw]
      • input/odometry [default: /localization/kinematic_state]
      • output/noise_tracks [default: noise_objects]
      • output/filtered_tracks [default: filtered_objects]
      • output/radar_detected_objects [default: detected_objects]
      • output/radar_tracked_objects [default: tracked_objects]
      • radar_tracks_msgs_converter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_msgs_converter.param.yaml]
      • radar_tracks_noise_filter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_noise_filter.param.yaml]

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged autoware_sensing_launch at Robotics Stack Exchange

No version for distro noetic showing github. Known supported distros are highlighted in the buttons above.
Package symbol

autoware_sensing_launch package from autoware_launch repo

autoware_launch autoware_sample_designs autoware_control_config autoware_control_launch autoware_localization_config autoware_localization_launch autoware_map_config autoware_map_launch autoware_perception_config autoware_perception_launch autoware_planning_config autoware_planning_launch autoware_sensing_launch autoware_sensor_driver_launch awsim_labs_sensor_kit_description awsim_labs_sensor_kit_launch awsim_sensor_kit_description awsim_sensor_kit_launch carla_sensor_kit_description carla_sensor_kit_launch sample_sensor_kit_description sample_sensor_kit_launch single_lidar_sensor_kit_description single_lidar_sensor_kit_launch tier4_autoware_api_launch tier4_simulator_launch tier4_system_launch tier4_vehicle_launch awsim_labs_vehicle_description awsim_labs_vehicle_launch sample_vehicle_description sample_vehicle_launch

ROS Distro
github

API Docs
Browse Code
Wiki

Package Summary

Version 0.53.0
License Apache License 2.0
Build type AMENT_CMAKE
Use RECOMMENDED

Repository Summary

Description
Checkout URI https://github.com/autowarefoundation/autoware_launch.git
VCS Type git
VCS Version main
Last Updated 2026-10-26
Dev Status UNKNOWN
Released UNRELEASED
Contributing Help Wanted (-)
Good First Issues (-)
Pull Requests to Review (-)

Package Description

Sensor data pre-processing pipelines shared by sensor kits

Maintainers

  • David Wong
  • Max Schmeller
  • Dai Nguyen

Authors

No additional authors.
autoware_universe_launch/autoware_sensing_launch/README.md

autoware_sensing_launch

Shared sensing launch library. Sensor kits own the sensing entry point (<sensor_model>_launch/launch/sensing.launch.xml, included by component_sensing.launch.xml); this package hosts the reusable pre/post-processing units the kits build on. Driver launch files live in the driver layer package autoware_sensor_driver_launch.

Container ownership and addressing

Sensor kits create the containers; units only load nodes into them. pointcloud_container.launch.xml is the one unit that creates a container, and it loads nothing into it.

Units that load composable nodes take the target container as an absolute path (target_container) supplied by the caller; units never derive it from their own namespace. Callers construct the path from their literal namespace knowledge and must keep it in step with their namespace pushes. The one exception is the concatenation units, whose pointcloud_container_name keeps its historical name and receives the top-level pointcloud container path passed through unchanged.

Launch files

  • pointcloud_container.launch.xml — component container for one sensor unit, created at container_namespace/container_name; use_multithread selects component_container_mt.
  • pointcloud_crop_boxes.launch.xml — vehicle body crop box followed by a secondary crop box, loaded into an existing container. Used on its own where the remaining preprocessing stages do not apply.
  • pointcloud_preprocessor_cpu.launch.xml — per-lidar preprocessing chain: the crop box unit, then distortion correction and ring outlier filter, loaded into an existing container. The secondary crop box is named by secondary_crop_box (e.g. mirror, wheels).
  • pointcloud_preprocessor_cuda.launch.xml — the same chain folded into one CUDA node; crop box stages come as folded lists in crop_box_param_path.
  • pointcloud_concatenation_cpu.launch.xml, pointcloud_concatenation_cuda.launch.xml — multi-lidar concatenation with time synchronization; the kit passes its concatenate_and_time_sync_node_param_path (input topic list is kit-specific).
  • blockage_diagnostics.launch.xml — blockage diagnostics on the raw pointcloud; angle_range defaults to the configured device angles (float-formatted).
  • gnss_postprocessor.launch.xml — septentrio heading converter plus NavSatFix-to-MGRS pose (autoware_gnss_poser).
  • imu_corrector.launch.xml — imu corrector plus gyro bias estimator (autoware_imu_corrector).
  • radar_tracks_postprocessor.launch.xml — radar tracks noise filter plus message converter.

Config

  • distortion_corrector_node.param.yaml, ring_outlier_filter_node.param.yaml — shared preprocessing defaults; kits with different sensors or thresholds pass their own file via the *_param_path arguments (the outlier shape parameters such as vertical_bins only take effect when publish_outlier_pointcloud is enabled).
  • blockage_diagnostics.param.yaml — blockage diagnostics defaults; runtime shape parameters (angle range, bins) are set per sensor via launch arguments.
  • radar_tracks_msgs_converter.param.yaml, radar_tracks_noise_filter.param.yaml — radar postprocessing defaults.

Crop box param files are owned by the sensor kits (they encode vehicle geometry). Kits generate them offline with:

scripts/generate_crop_box_params.py <vehicle_info.param.yaml> <mirror.param.yaml> <output_dir>

and pass the outputs via the crop_box_filter_{self,secondary}_param_path arguments.

Scripts

  • septentrio_heading_converter.py — runtime node converting the septentrio AttEuler heading to a GnssInsOrientationStamped (installed).
  • generate_crop_box_params.py — offline generator for kit crop box files (not installed).

Package Dependencies

Please see <exec_depend> in package.xml.

CHANGELOG

Changelog for package autoware_sensing_launch

0.53.0 (2026-10-13)

  • Merge remote-tracking branch 'origin/main' into tmp/bot/bump_version_base

  • feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled (#1987)

    * feat(autoware_sensing_launch): run blockage_diag standalone when agnocast is enabled autoware_universe#13074 moves BlockageDiagComponent to agnocast_wrapper::Node, which is not designed to run inside a component container. Start the node as a standalone process when ENABLE_AGNOCAST=1 and keep loading it into target_container otherwise, as component_state_monitor.launch.py does for topic_state_monitor. The node name, namespace, remaps and parameters are identical in both branches.

    * refactor(autoware_sensing_launch): launch blockage_diag through the autoware_node action ---------

  • feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled (#1990)

    * feat(autoware_launch): run crop_box_filter standalone when agnocast is enabled autoware_universe#13384 moves CropBoxFilterComponent to agnocast_wrapper::Node on an agnocast-only executor, which the component containers do not set up. Start each of the eight crop_box_filter nodes as a standalone process when ENABLE_AGNOCAST=1 and keep loading them into their container otherwise. The node names, namespaces, remaps and parameters are the same in both branches.

    * refactor(autoware_launch): launch crop_box_filter through the autoware_node action ---------

  • refactor(sensing)!: separate sensor driver and pre-processing layers from sensor kit configuration (#1938)

    * refactor(launch)!: rename tier4_sensing_launch to autoware_sensing_launch tier4_sensing_launch held one 19-line pass-through page that pushed the sensing namespace and included the sensor kit entry. The component page now includes that entry directly and pushes the namespace itself, so the layer is gone and the package is left as the home of the shared sensing units added next.

    • move the package to autoware_universe_launch/autoware_sensing_launch

    - rename tier4_sensing_component.launch.xml to components/component_sensing.launch.xml

    - drop the pass-through page and its diagram Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensing_launch): add the shared sensing unit library Declarative XML units for the sensor data pre/post-processing stages, each addressed by an absolute container path so a caller can place it in any container it owns.

    • pointcloud_container, the only unit that creates a container

    - pointcloud_crop_boxes, pointcloud_preprocessor_{cpu,cuda}, pointcloud_concatenation_{cpu,cuda}

    - blockage_diagnostics, gnss_postprocessor, imu_corrector, radar_tracks_postprocessor

    • shared preprocessing, blockage and radar defaults under config/

    - generate_crop_box_params.py, which turns vehicle info and mirror geometry into the crop box files the kits commit and pass in; septentrio_heading_converter.py is installed as a runtime node The parameter files and the heading converter are ported from tier4/aip_launcher (aip_common_sensor_launch), Apache-2.0 under the same TIER IV copyright. Nothing includes these units yet. Co-Authored-By: Claude Opus 5 (1M context) <<noreply@anthropic.com>>

    * feat(autoware_sensor_driver_launch): add the sensor driver layer The hardware interface layer, kept apart from the pre-processing units so that a sensor make and model can be brought into the ROS graph on its own.

    - nebula_{velodyne,robosense,hesai}: per-make driver chains, configured by the model param file shipped in the nebula driver package

    - velodyne_{VLP16,VLS128}, robosense_{Bpearl,Helios}, hesai_{XT32,OT128}: per-model wrappers

    - every file loads its node into target_container, an existing container owned by the caller; this package never creates one The

File truncated at 100 lines see the full file

Launch files

  • launch/blockage_diagnostics.launch.xml
      • target_container
      • node_name [default: blockage_diag]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • output/pointcloud [default: blockage_diag/pointcloud]
      • cloud_min_angle [default: 0.0]
      • cloud_max_angle [default: 360.0]
      • angle_range [default: [$(var cloud_min_angle), $(var cloud_max_angle)]]
      • max_range [default: 300.0]
      • horizontal_ring_id [default: 64]
      • vertical_bins [default: 128]
      • is_channel_order_top2down [default: true]
      • horizontal_resolution [default: 0.4]
      • blockage_diagnostics_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/blockage_diagnostics.param.yaml]
  • launch/gnss_postprocessor.launch.xml
      • input/fix [default: septentrio/nav_sat_fix]
      • input/atteuler [default: septentrio/atteuler]
      • input/orientation [default: septentrio/orientation]
      • output/pose [default: pose]
      • output/pose_with_covariance [default: pose_with_covariance]
      • output/fixed [default: fixed]
  • launch/imu_corrector.launch.xml
      • input/imu_raw
      • output/imu_data [default: imu_data]
      • input/odometry [default: /localization/kinematic_state]
      • input/pose_with_covariance [default: /localization/pose_estimator/pose_with_covariance]
      • output/imu_scaled [default: /sensing/imu_output_scaled]
      • imu_corrector_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/imu_corrector.param.yaml]
      • gyro_scale_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_scale_estimator.param.yaml]
      • gyro_bias_estimator_param_path [default: $(find-pkg-share autoware_imu_corrector)/config/gyro_bias_estimator.param.yaml]
  • launch/pointcloud_concatenation_cpu.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • use_intra_process [default: false]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_concatenation_cuda.launch.xml
      • pointcloud_container_name [default: pointcloud_container]
      • use_concat_filter [default: true]
      • concatenate_and_time_sync_node_param_path
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • output/pointcloud [default: concatenated/pointcloud]
      • output/pointcloud_info [default: concatenated/pointcloud_info]
  • launch/pointcloud_container.launch.xml
      • container_namespace
      • container_name [default: pointcloud_container]
      • use_multithread [default: false]
  • launch/pointcloud_crop_boxes.launch.xml
      • target_container
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • intermediate/pointcloud [default: self_cropped/pointcloud_ex]
      • output/pointcloud [default: $(var secondary_crop_box)_cropped/pointcloud_ex]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
  • launch/pointcloud_preprocessor_cpu.launch.xml
      • target_container
      • frame_id [default: lidar]
      • input_frame [default: base_link]
      • output_frame [default: base_link]
      • output_as_sensor_frame [default: true]
      • secondary_crop_box [default: mirror]
      • use_intra_process [default: false]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_filter_self_param_path
      • crop_box_filter_secondary_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/pointcloud_preprocessor_cuda.launch.xml
      • target_container
      • enable_ring_outlier_filter [default: true]
      • input/pointcloud [default: pointcloud_raw_ex]
      • input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
      • input/imu [default: /sensing/imu/imu_data]
      • output/pointcloud [default: pointcloud_before_sync]
      • crop_box_param_path
      • distortion_corrector_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/distortion_corrector_node.param.yaml]
      • ring_outlier_filter_node_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/ring_outlier_filter_node.param.yaml]
  • launch/radar_tracks_postprocessor.launch.xml
      • input/tracks [default: objects_raw]
      • input/odometry [default: /localization/kinematic_state]
      • output/noise_tracks [default: noise_objects]
      • output/filtered_tracks [default: filtered_objects]
      • output/radar_detected_objects [default: detected_objects]
      • output/radar_tracked_objects [default: tracked_objects]
      • radar_tracks_msgs_converter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_msgs_converter.param.yaml]
      • radar_tracks_noise_filter_param_path [default: $(find-pkg-share autoware_sensing_launch)/config/radar_tracks_noise_filter.param.yaml]

Messages

No message files found.

Services

No service files found

Plugins

No plugins found.

Recent questions tagged autoware_sensing_launch at Robotics Stack Exchange