ROS Index
Package Summary
| Tags | No category tags. |
| Version | 0.46.0 |
| License | Apache License 2.0 |
| Build type | AMENT_CMAKE |
| Use | RECOMMENDED |
Repository Summary
| Description | |
| Checkout URI | https://github.com/autowarefoundation/autoware_universe.git |
| VCS Type | git |
| VCS Version | main |
| Last Updated | 2025-07-30 |
| Dev Status | UNKNOWN |
| CI status | No Continuous Integration |
| Released | UNRELEASED |
| Tags | planner ros calibration self-driving-car autonomous-driving autonomous-vehicles ros2 3d-map autoware |
| Contributing |
Help Wanted (0)
Good First Issues (0) Pull Requests to Review (0) |
Package Description
Additional Links
Maintainers
- Kenzo Lobos-Tsunekawa
- Amadeusz Szymko
- Max Schmeller
- David Wong
- Manato Hirabayashi
- Yihsiang Fang
Authors
- Kenzo Lobos-Tsunekawa
autoware_cuda_pointcloud_preprocessor
Purpose
The pointcloud preprocessing implemented in autoware_pointcloud_preprocessor has been thoroughly tested in autoware. However, the latency it introduces does not scale well with modern LiDAR devices due to the high number of points they introduce.
To alleviate this issue, this package reimplements most of the pipeline presented in autoware_pointcloud_preprocessor leveraging the use of GPGPUs. In particular, this package makes use of CUDA to provide accelerated versions of the already established implementations, while also maintaining compatibility with normal ROS nodes/topics. <!– cSpell: ignore GPGPUs >
Inner-workings / Algorithms
A detailed description of each filter’s algorithm is available in the following links.
| Filter Name | Description | Detail |
|---|---|---|
| cuda_pointcloud_preprocessor | Implements the cropping, distortion correction, and outlier filtering (ring-based) of the autoware_pointcloud_preprocessor’s CPU versions. |
link |
| cuda_concatenate_and_time_sync_node | Implements pointcloud concatenation an synchronization following autoware_pointcloud_preprocessor’s CPU implementation. |
link |
| cuda_voxel_grid_downsample_filter | Implements voxel downsample filtering of the autoware_pointcloud_preprocessor’s CPU version |
link |
(Optional) Future extensions / Unimplemented parts
The subsample filters implemented in autoware_pointcloud_preprocessor will have similar counterparts in this package.
Changelog for package autoware_cuda_pointcloud_preprocessor
0.46.0 (2025-06-20)
- Merge remote-tracking branch 'upstream/main' into tmp/TaikiYamada/bump_version_base
- feat(autoware_cuda_pointcloud_preprocessor): agnocast support
(#10812)
- feat(autoware_cuda_pointcloud_preprocessor): add Agnocast support for incoming pointclouds
- chore: make compilable both with and without agnocast
- style(pre-commit): autofix
* ci: statisfy cppcheck and cmake_lint ---------Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]\@users.noreply.github.com>
- feat(autoware_cuda_pointcloud_preprocessor): diagnostic for cuda
pointcloud preprocessor
(#10793)
- feat: add ring, crop box diag
- feat: add distortion correction
- feat: add concat diagnostics
- chore: remove if debug publisher
- chore: clean code
- chore: clean code
- chore: utilize mask instead of atomicadd
- chore: count nan points and numbers of point after crop box filter
- chore: fix schema
- chore: move to structure
- chore: update library
- chore: prefix output
- chore: chagne shared pointer to const reference
- chore: use device vector for thrust count
- chore: fix output pointcloud name
- chore: add comment
- chore: reuse function
- chore: fix merging issue
- chore: prefix output
- chore: fix layout
- chore: add doc comment
* chore(autoware_cuda_pointcloud_preprocessor): disable uncrustify ---------Co-authored-by: Max SCHMELLER <<max.schmeller@tier4.jp>>
- chore(autoware_cuda_pointcloud_preprocessor): add myself as maintainer (#10809)
- fix(cuda_pointcloud_preprocessor): ensure ordered twist/imu queues
(#10748)
- fix(cuda_pointcloud_preprocessor): ensure ordered twist/imu queues
- chore: satisfy uncrustify
* chore: uncrustify and clang-format conflict, disable uncrustify for statement ---------Co-authored-by: Max SCHMELLER <<msc.schmeller@tier4.jp>>
- feat(cuda_pointcloud_preprocessor): update filtering parameter and process (#10555)
- fix(cuda_pointcloud_preprocessor): reset data when receiving zero
siz…
(#10723)
- fix(cuda_pointcloud_preprocessor): reset data when receiving zero size pointcloud
- fix(cuda_pointcloud_preprocessor): hotfix for ghost output
- insert memory region reset for every iteration
- judge if each CUDA thread treat valid input point (or not)
- fix(cuda_pointcloud_preprocessor): apply valid point mask
* style(pre-commit): autofix ---------Co-authored-by: Manato HIRABAYASHI <<manato.hirabayashi@tier4.jp>> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]\@users.noreply.github.com> Co-authored-by: Manato Hirabayashi <<3022416+manato@users.noreply.github.com>>
- fix(autoware_cuda_pointcloud_preprocessor): fix CMakeLists.txt to install cuda_pointcloud_preprocessor library (#10740) Co-authored-by: Amadeusz Szymko <<amadeusz.szymko.2@tier4.jp>>
- feat: accelerate voxel filter
(#10566)
- feat: add cuda_voxel_grid_downsample_filter
- refactor(cuda_voxel_grid_downsample_filter): clean up codes
- fix(cuda_voxel_grid_downsample_filter): suppress warning for arithmetic on pointer to void
- chore(cuda_voxel_grid_dowmsample_filter): remove debug code
* fix(cuda_voxel_grid_downsample_filter): support XYZIRC output format Set output format to [Cloud XYZIRC]{.title-ref} according to the [design document](https://autowarefoundation.github.io/autoware-documentation/main/design/autoware-architecture/sensing/data-types/point-cloud/)
- style(pre-commit): autofix
- fix(cuda_boxel_grid_downsamle_filter): rearrange package structure
- chore(cuda_voxel_grid_dowmsample_filter): reuse OutputPointType in parent namespace
- chore(cuda_voxel_grid_downsample_filter): use macro defined in autoware_cuda_utils for error checking
- feat(cuda_voxel_grid_downsample_filter): support multiple data types for input intensity
- fix(cuda_voxel_grid_downsample_filter): cleanup included header files
* chore: correct comments Co-authored-by: Kenzo Lobos Tsunekawa <<kenzo.lobos@gmail.com>>
- style(pre-commit): autofix
- feat(cuda_voxel_grid_downsample_filter): use cub instead of thrust for better acceleration
- style(pre-commit): autofix
- feat(cuda_voxel_grid_downsample_filter): use dedicated memory pool
- feat(cuda_voxel_grid_downsample_filter): introduce a parameter to control max size for GPU memory pool
- docs: add/modify schema and documents for cuda_voxel_grid_downsample_filter
- style(pre-commit): autofix
- chore: fix spell miss
- refactor: fix code style divergence error
- style(pre-commit): autofix
- fix: re-add INDENT-ON/OFF
* feat: use most significant bit calculation to make radix sort faster ---------Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]\@users.noreply.github.com> Co-authored-by: Kenzo Lobos Tsunekawa <<kenzo.lobos@gmail.com>>
- Contributors: Fumiya Watanabe, Kotaro Uetake, Manato Hirabayashi, Max Schmeller, TaikiYamada4, Yi-Hsiang Fang (Vivid), keita1523
0.45.0 (2025-05-22)
- Merge remote-tracking branch 'origin/main' into tmp/notbot/bump_version_base
- feat(autoware_cuda_pointcloud_preprocessor): added target
architectures for the cuda pointcloud preprocessor
(#10612)
- chore: added target architectures for the cuda pointcloud preprocessor
- chore: mistook the compute capabilities of edge devices
* chore: cspell ---------
- perf(autoware_tensorrt_common): set cudaSetDeviceFlags explicitly
(#10523)
- Synchronize CUDA stream by blocking instead of spin
- Use blocking-sync in BEVFusion
- Call cudaSetDeviceFlags in tensorrt_common
- feat(autoware_cuda_pointcloud_preprocessor): replace imu and
twist callback with polling subscriber
(#10509)
- feat(cuda_pointcloud_preprocessor): replace subscriptions with InterProcessPollingSubscriber for twist and IMU data
- fix(cuda_pointcloud_preprocessor): remove unused twist_queue_ variable
* style(pre-commit): autofix ---------Co-authored-by: Takahisa.Ishikawa <<takahisa.ishikawa@tier4.jp>> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]\@users.noreply.github.com> Co-authored-by: Kenzo Lobos Tsunekawa <<kenzo.lobos@tier4.jp>>
- feat(autoware_cuda_pointcloud_preprocessor): pointcloud
concatenation
(#10300)
- feat: cuda accelerated version of the pointcloud concatenation
- chore: removed duplicated include
- chore: changed to header blocks from pragmas :c
- chore: removed yaml and schema since this node uses the same interface as the non-gpu node
- chore: fixed rebased induced error
- fix: used the wrong point type
- chore: changed pointer to auto
- chore: rewrote equation for clarity
- chore: added a comment regarding the reallocation strategy
- chore: reflected latest changes in the templated version of the concat
- chore: addressed cppcheck reports
- chore: fixed dead link
- chore: solving uncrustify conflicts
- chore: more uncrustify
- chore: yet another uncrustify related error
- chore: hopefully last uncrustify error
* chore: now fixing uncrustify on source files ---------
- Contributors: Kenzo Lobos Tsunekawa, TaikiYamada4, Takahisa Ishikawa, prime number
0.44.2 (2025-06-10)
0.44.1 (2025-05-01)
0.44.0 (2025-04-18)
0.43.0 (2025-03-21)
- fix: update tool version
- Merge remote-tracking branch 'origin/main' into chore/bump-version-0.43
- chore(autoware_cuda_pointcloud_preprocessor): add maintainer (#10297)
- feat(autoware_cuda_pointcloud_preprocessor): a cuda-accelerated
pointcloud preprocessor
(#9454)
- feat: moved the cuda pointcloud preprocessor and organized from a personal repository
- chore: fixed incorrect links
- chore: fixed dead links pt2
- chore: fixed spelling errors
- chore: json schema fixes
- chore: removed comments and filled the fields
- fix: fixed the adapter for the case when the number of points in the pointcloud changes after the first iteration
- feat: used the cuda host allocators for aster host to device copies
* Update sensing/autoware_cuda_pointcloud_preprocessor/docs/cuda-pointcloud-preprocessor.md Co-authored-by: Max Schmeller <<6088931+mojomex@users.noreply.github.com>> * Update sensing/autoware_cuda_pointcloud_preprocessor/src/cuda_pointcloud_preprocessor/cuda_pointcloud_preprocessor.cu Co-authored-by: Manato Hirabayashi <<3022416+manato@users.noreply.github.com>> * Update sensing/autoware_cuda_pointcloud_preprocessor/src/cuda_pointcloud_preprocessor/cuda_pointcloud_preprocessor.cu Co-authored-by: Manato Hirabayashi <<3022416+manato@users.noreply.github.com>>
- style(pre-commit): autofix
* Update sensing/autoware_cuda_pointcloud_preprocessor/docs/cuda-pointcloud-preprocessor.md Co-authored-by: Max Schmeller <<6088931+mojomex@users.noreply.github.com>> * Update sensing/autoware_cuda_pointcloud_preprocessor/README.md Co-authored-by: Max Schmeller <<6088931+mojomex@users.noreply.github.com>> * Update sensing/autoware_cuda_pointcloud_preprocessor/README.md Co-authored-by: Max Schmeller <<6088931+mojomex@users.noreply.github.com>> * Update sensing/autoware_cuda_pointcloud_preprocessor/src/cuda_pointcloud_preprocessor/cuda_pointcloud_preprocessor.cu Co-authored-by: Max Schmeller <<6088931+mojomex@users.noreply.github.com>>
- style(pre-commit): autofix
* Update sensing/autoware_cuda_pointcloud_preprocessor/src/cuda_pointcloud_preprocessor/cuda_pointcloud_preprocessor.cu Co-authored-by: Manato Hirabayashi <<3022416+manato@users.noreply.github.com>>
- style(pre-commit): autofix
* Update sensing/autoware_cuda_pointcloud_preprocessor/src/cuda_pointcloud_preprocessor/cuda_pointcloud_preprocessor.cu Co-authored-by: Manato Hirabayashi <<3022416+manato@users.noreply.github.com>>
- style(pre-commit): autofix
- chore: fixed code compilation to reflect Hirabayashi-san's memory pool proposal
- feat: generalized the number of crop boxes. For two at least, the new approach is actually faster
- chore: updated config, schema, and handled the null case in a specialized way
- feat: moving the pointcloud organization into gpu
- feat: reimplemented the organized pointcloud adapter in cuda. the only bottleneck is the H->D copy
- chore: removed redundant ternay operator
- chore: added a temporary memory check. the check will be unified in a later PR
- chore: refactored the structure to avoid large files
- chore: updated the copyright year
- fix: fixed a bug in the undistortion kernel setup. validated it comparing it with the baseline
- chore: removed unused packages
- chore: removed mentions of the removed adapter
- chore: fixed missing autoware prefix
- fix: missing assignment in else branch
- chore: added cuda/nvcc debug flags on debug builds
- chore: refactored parameters for the undistortion settings
- chore: removed unused headers
- chore: changed default crop box to no filtering at all
- feat: added missing restrict keyword
- chore: spells
- chore: removed default destructor
- chore: ocd activated (spelling)
- chore: fixed the schema
- chore: improved readibility
- chore: added dummy crop box
- chore: added new repositories to ansible
- chore: CI/CD
- chore: more CI/CD
- chore: mode CI/CD. some linters are conflicting
- style(pre-commit): autofix
- chore: ignoring uncrustify
- chore: ignoring more uncrustify
- chore: missed one more uncrustify exception
* chore: added meta dep ---------Co-authored-by: Max Schmeller <<6088931+mojomex@users.noreply.github.com>> Co-authored-by: Manato Hirabayashi <<3022416+manato@users.noreply.github.com>> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]\@users.noreply.github.com> Co-authored-by: Amadeusz Szymko <<amadeusz.szymko.2@tier4.jp>>
- Contributors: Amadeusz Szymko, Hayato Mizushima, Kenzo Lobos Tsunekawa
Wiki Tutorials
Package Dependencies
System Dependencies
Dependant Packages
Launch files
- launch/cuda_voxel_grid_downsample_filter.launch.xml
-
- input/pointcloud [default: /sensing/lidar/top/pointcloud_raw]
- output/pointcloud [default: /sensing/lidar/top/downsampled]
- debug [default: false]
- cuda_voxel_gird_downsample_filter_param_file [default: $(find-pkg-share autoware_cuda_pointcloud_preprocessor)/config/cuda_voxel_grid_downsample_filter.param.yaml]
- launch/cuda_pointcloud_preprocessor.launch.xml
-
- input/pointcloud [default: /sensing/lidar/top/pointcloud_raw_ex]
- input/imu [default: /sensing/imu/imu_data]
- input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
- output/pointcloud [default: /sensing/lidar/top/test]
- cuda_pointcloud_preprocessor_param_file [default: $(find-pkg-share autoware_cuda_pointcloud_preprocessor)/config/cuda_pointcloud_preprocessor.param.yaml]
- launch/cuda_concatenate_and_time_sync_node.launch.xml
-
- input/twist [default: /sensing/vehicle_velocity_converter/twist_with_covariance]
- output [default: /sensing/lidar/concatenated/pointcloud]
- output_info [default: /sensing/lidar/concatenated/pointcloud_info]
- param_file [default: $(find-pkg-share autoware_pointcloud_preprocessor)/config/concatenate_and_time_sync_node.param.yaml]