support dynamic composed bringup #2750
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
|
@gezp, please properly fill in PR template in the future. @SteveMacenski, use this instead.
|
SteveMacenski
requested changes
Jan 3, 2022
gezp
force-pushed
the
dynamic_composed_bringup
branch
from
b055103
to
ba23877
Compare
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
gezp
force-pushed
the
dynamic_composed_bringup
branch
from
ba23877
to
26493eb
Compare
|
updated performance test
i re-run all test cases, the performance of dynamic composition which uses btw, the cpu is measured by |
|
@gezp, your PR has failed to build. Please check CI outputs and resolve issues. |
|
Interesting, do you know why it is that manual composition isn't as good as dynamic composition? I would have assumed the opposite |
current manual composition in nav2 only composes navigation nodes, and localization nodes are still normal multiple processes, so manual composition consumes higher memory than dynamic composition. in terms of cpu, manual composition and dynamic composition (use |
|
Got it, something to note and make sure we find a way around for experiments later for the paper. We should make sure they're as equivalent as possible. e.g. we have a Please make the changes to the code I requested above and I can approve this and we can merge it in once there's a rolling sync. This is a really big contribution @gezp ! Thanks for sticking with it and making contributions across the ecosystem. This is definitely a long term impactful project. |
|
I believe rolling was updated! |
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
|
yes, i can use |
SteveMacenski
approved these changes
Jan 20, 2022
|
Anything else we need to consider before merging in? Please submit a PR to the documentation website to the migration guide, somewhere near the top (https://navigation.ros.org/migration/Galactic.html) talking about changing to dynamic composition and the performance improvement it represents (a table or something would be nice showing the performance achieved) . |
|
@gezp I believe all that's blocking this is the documentation PR that should be very short using metrics you've already collected. |
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
|
i created a PR for migration docs in |
SteveMacenski
approved these changes
Jan 26, 2022
SteveMacenski
approved these changes
Jan 26, 2022
sergigraum
pushed a commit
to umdlife/navigation2
that referenced
this pull request
Aug 26, 2022
* Name bt_navigator action nodes uniquely (#2410)
Co-authored-by: Jonatan Olofsson <jonatan.olofsson@saabgroup.com>
* Fix code coverage (#2419)
* Include filter only by packge name
as using package source paths can exclude coverage from other packages
such as header files reported usign under workspace's install path
* formating
* Process gcno
Process both gcno and gcda coverage files.
This option is useful for capturing untested files in the coverage report.
* Validate sources
Check if every source file exists
* Dump statistic
Dump total statistic at the end
* Branch coverage
Include only the most useful branches in the coverage report.
* Fix install path
* Explicitly exclude packages
to ignore rviz package
but still explicitly include packages
as the include filter is what mostly speeds up processing
* Revert "Branch coverage"
This reverts commit 20f28949904ef965c8b5b011f3e387037adc93ce.
* Keep path fixes relative to workspace
As the working directory used for running code coverage report script
is already at the root of the colcon workspace,
avoiding the use of unnecessary path globbing
* Followup to PR 2413 for main branch (fix for operator= in observation.hpp) (#2425)
* in observation.hpp fixed operator=
* style fix
* Update README.md
* Add lazy_theta_star (#1839)
* Add lazy_theta_star
* Added license
* Added license
* Update lazy_theta_star_b.cpp
* restructured files
- separated planner part and the algorithm
- implemented all the changes as suggested
* restructured files
* Update CMakeLists.txt
* removed unnecessary comments
* removed comments
* Delete planner.cpp
* Delete lazy_theta_star_b.cpp
* Delete lazy_theta_star_b.h
* Delete planner.h
* Delete CMakeLists.txt
* Delete global_planner_plugin.xml
* Delete package.xml
* replaced the files
- refactored code
- improved reliability
- have to write a code similar to that in nav2_system_tests, to test it (working on it)
* removed comments
* removed comments
* Update lazy_theta_star2.cpp
* update files
- replaced manual management of priority queue with stl priority queue
- added the parameter ".lethal_cost"
- removed unnecessary parameters passed to the functions
* update files
- updated the header files in accordance to their .cpp counterparts
* Delete lazy_theta_star2.h
* Delete lazy_theta_star_planner.h
* Delete lazy_theta_star2.cpp
* Delete lazy_theta_star_planner.cpp
* Delete CMakeLists.txt
* Delete package.xml
* Delete global_planner_plugin.xml
* upload the changed code
Changes from last time are:
- the code has been changed to the Lazy Theta* P variant, in order to account for the costmap traversal costs
- parameters are available to change the weights of the costmap traversal cost (weight = 1.75, as of now) and the distance function (weight = 1.0, as of now
-
* Delete lazy_theta_star_p_planner directory
* Replace the old files
- the structure of code has been changed
- new functions have been added, namely : getTraversalCost, getEuclideanCost, getCellCost, isSafe[it is now an overloaded function]
- documentation added for variables and functions
- the parameters for the planner now consists of : how_many_corners, costmap_tolerance, euc_tolerance (documentation to added soon)
- fixed a bug where the incorrect traversal cost of the node was taken
* Delete lazy_theta_star_p_planner directory
* update the files
- renamed the project to nav2_theta_star_planner from lazy_theta_star_p_planner
- renamed files from lazy_theta_p_planner.hpp/.cpp to theta_star_planner.hpp/.cpp and lazy_theta_star.hpp/.cpp to theta_star.hpp/.cpp
- added a readme file outlining the parameters, usage notes and images to be added soon
- added parameters and renamed the parameters for the cost function (costmap_tolerance -> w_traversal_cost ; euc_tolerance -> w_euc_cost ; added a parameter for the heuristic)
- replaced the SharedPtr with a WeakPtr for node
- removed +1 and the pusher_ variable added to compensate for it
* Update README.md
* update the code
- linted the code
- **updates to readme, are still pending**
- changed the type of message from INFO to DEBUG
- replaced the capital letters with the smaller ones
* update the readme file
* Update README.md
* Update README.md
* Update README.md
* Update README.md
* Delete global_planner_plugin.xml
* fix the linting issues
* remove the space on line 7
* change RCLCPP_INFO to RCLCPP_DEBUG
* Update README.md
* Add test coverage
- added tests to check the algorithm itself and its helper functions
- added smoke test to detect plugin level issues
- inlined some functions
- shifted the functions `setStartAndGoal()` and `isSafeToPlan()` to the ThetaStar class
- removed the functions `dist()` and `getCellCost()` from the ThetaStar class
* update the test file
* update the test file
* Update README.md
* Update README.md
* Update README.md
* add test on the size of the output path
* Update README.md
* Update README.md
* change the function name from `isSafeToPlan()` to `isUnsafeToPlan()`
* update the files
- inlined the functions `getIndex` and `addIndex`
- removed typos from the comments
* fix typos
* Update README.md
* Update README.md
* Update README.md
* Update README.md
* Update README.md
* Update README.md
* Update README.md
* fix the typo
- the first isSafe calls output wasn't negated, it has been fixed
* Update theta_star.hpp
- added the `getCost()` function (again)
- replaced the use of indices with pointers to store the node's data, changes were made to the following functions - `addIndex()`, `getIndex()`, `initialisePosn()`.
- the priority queue now stores the pointer to node's data and accordingly changes were made to `comp` struct
- the global variable `curr_node` was renamed to `exp_node`
- removed the struct `pos`
* Update theta_star.cpp
* update default parameters
* update default parameters
* Update README.md
* update the test file
* fix linting issues
* Update README.md
* Update README.md
* Update README.md
* Update theta_star_planner.cpp
* update LETHAL_COST
* update README.md
* update dependency list
* Update package.xml
* Remove deprecated pluginlib boost feature disable (#2431)
Boost features have been deprecated and removed
in pluginlib for rolling
* Improve coverage for NavFn package (#2420)
* feat(navfn): improve coverage
* wip(navfn): add new test launch
* refactor(navfn): check expect_failure flag value
* Reduce nodes for nav2_waypoint_follower (#2441)
* Reduce node of waypoint_follower by using callback group and executor
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* fix linting failures
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* missing comma in bt_navigator plugin list (#2447)
* Feature addition: capability for the RRP to drive the robot backwards (#2443)
* Feature addition: capability for the RRP to drive the robot backwards
* findDirectionChange(): returns maximum double value + removing the size check for the indexer in for loop
* satisfying ament_uncrustify and ament_cpplint
* removing the unnecessary condition in the cusp determination
* Satisfying ament_uncrustify - 2
* Adding launch_testing_ros dep on nav2 utils to install (#2450)
* Update README.md (#2461)
* Refactor actions to rebuild CI image from CI (#2405)
* Add dependabot config
* Workarround missing ARG support in dependabot
Parrent image can't yet be parameterized for version parsing
https://github.com/dependabot/dependabot-core/issues/2057
* Fix CCACHE_DIR to be absolute path
https://github.com/ros-planning/navigation2/pull/2403
* Remove FAIL_ON_BUILD_FAILURE ARG
now that CI image builds from rolling testing
and no longer ros2 nightly
so forcing image rebuilds is no longer needed
* Trigger rebuild on changes to Dockerfile
* Add github-actions to dependabot
* Refactor actions to rebuild CI image from CI
* Clean old dockerhub build hook
* Rename action
* Combine workflows to avoid dispaching
as that requires a github personal access token
which is an inconvenience to generate and add to repo secrets
* Add needs param to define job order
* Use job outputs for condition instead of env
Seems that env in context is not yet avalable at the job level
but only at the current job's step level
https://github.community/t/how-to-set-and-access-a-workflow-variable/17335/6
https://docs.github.com/en/actions/reference/workflow-syntax-for-github-actions#jobsjob_idoutputs
* Use multi-line syntax for tags param
* Use toJSON for job conditional
* Test with always
* Update param syntax
* Test with always
* Update if syntax
* Test if trigger for only check_ci_files
* Simplify if params
* Try using GitHub's Container registry
* Indent formating
* Add step conditional
* Default arg to build from unpinned parrent stage
allowing CI to build from pinned testing image
while enabling dependabot to find and update testing image version
ensuring CI images are up-to-date with upstream registries
* Set build arg for continuous integration
* Ignore .github folder
* testing
* Update colcon-cache
* Revert to rolling
* Revert testing
* Update README
* SMAC Planner: Treating "inscribed" costs not as collision anymore (#2464)
* Reduce map saver nodes (#2454)
* reduce MapSaver nodes by using callback group/executor combo
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* set use_rclcpp_node false
* a cleaner solution using a future and spin_until_future_complete()
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* Update nav2_controller.cpp (#2462)
Add `costmap_thread_.reset()` on the destructor of ControllerServer class
* Reduce lifecycle manager nodes (#2456)
* remove bond_client_node_ in class LifecycleManager
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* clear unused variables
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* fix lint
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* use dedicated executor thread
Signed-off-by: William Woodall <william@osrfoundation.org>
* fix building error
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* support to process executor for NodeThread
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* use NodeThread for LifecycleManager
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
Co-authored-by: William Woodall <william@osrfoundation.org>
* sync with main and use ros_diff_drive in case name changes (#2472)
Co-authored-by: YOUSSEF LAHROUNI <youssef.lahrouni@proioxrobots.com>
* Reduce lifecycle service client nodes (#2469)
* remove LifecycleServiceClient's constructor which create internal node
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* fix the linter
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* Python string format (#2466)
* Convert to python format strings for readability
* Merge concatenated strings
* Revert converting generated files
* Single quotes for consistency
* Just print the exception
* Fix Smac cleanup (#2477)
* fix smac2d cleanup
* same for hybrid
* Naming BT client node after action name (#2470)
* Put action name in node namespace instead of node name
* Put namespace remapping in node options
* Rename client node with action name replacing "/" by "_"
* Code formatting
* Code formatting
* improve SimpleActionServer (#2459)
* improve SimpleActionServer
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
add a bool argument spin_thread whether use callback group/executor combo with dedicated thread to handle server
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* use Nodethread for executor
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* add constructor's arg options and member variable executor_
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* fix nav2 params and launch file to publish Local and global costmaps in multi robots example (#2471)
* adjust launch file with needed gazebo plugin and set groot to False for multi-robot params
* correct unwanted changes
* change port and set groot to false
* fix lints
Co-authored-by: YOUSSEF LAHROUNI <youssef.lahrouni@proioxrobots.com>
* reduce planner server nodes (#2480)
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* Reduce controller server nodes (#2479)
* reduce controller server nodes by setting SimpleActionServer's argument spin_thread true
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* remove duplicated action_server_.reste()
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* Update regulated_pure_pursuit_controller.cpp (#2484)
* [SmacPlanner2D] make tolerance parameter dynamic (#2475)
* make tolerance dyn param
* full reconfigure
* fix typo
* Place function above the variables
* lock param callback
* Modify the BtServiceNode to include an on_success call. (#2481)
* Modify the BtServiceNode to include an on_success call.
* PR: Fix linter error by removing trailing whitespaces.
* PR: Rename on_success() to on_completion() to improve understandability.
* Accept path of length 1 (#2495)
* Use regex anchors when checking for updates (#2500)
to fix conner cases like this:
https://github.com/ros-planning/navigation2/runs/3277796714?check_suite_focus=true
https://stackoverflow.com/a/4709921
* Update README.md
* adding default template constructor (#2501)
* Fix null pointer in amcl on_cleanup (#2503)
* fix data race: addPlugin() and resizeMap() can be executed concurrently (#2512)
Co-authored-by: Kai-Tao Xie <kaitaoxie@qq.com>
* fix data race: VoxelLayer::matchSize may be executed concurrently (#2513)
Co-authored-by: Kai-Tao Xie <kaitaoxie@qq.com>
* catch runtime_error if the message from laser is malformed (#2511)
* catch runtime_error if the message from laser is malformed
* fix styling
Co-authored-by: Kai-Tao Xie <kaitaoxie@qq.com>
* Smac planner bad alloc (#2516)
* test(nav2_smac_planner): show short path bad_alloc
When given a goal that is one or zero costmap
cells away, AStarAlgorithm throws std::bad_alloc
* fix(nav2_smac_planner): fixed bad_alloc
* [ObstacleLayer] Use message_filter timeout (#2518)
* , tf2::durationFromSec(transform_tolerance)
* use message_filter timeout in AMCL
* also for sensor_msgs::msg::PointCloud2
* fix possible use-after-free: unsafe shared_ptr in multithread (#2510)
Co-authored-by: Kai-Tao Xie <kaitaoxie@qq.com>
* Spawn turtlebot3 separately (#2473)
* add nav2 gazebo spawner
* fix origin pose
* add robot_description
* remove urdf arg
* fix whitespace
* spawn from file ;add initial pose
* remove spawn launch include
* param and rviz edit
* conflict resolve
* pose
* remove world tags and add tf remap
* use nav2 waffle model and spawn entity
* change nav2_spawner to spawn_entity
* add tf remap
* delete nav2_gazebo_spawner
* sdf param
* added in readme
* changes
* fix export dependency and library (#2521)
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* Add argument node options (#2522)
* add argument node_options
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* change node's name
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* fix code style
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* Add more semantic checks for amcl parameters (#2528)
* Fix null pointer in amcl on_cleanup
* Add more semantic checks for amcl
* fix possible use-after-free: unsafe shared_ptr in multithread (#2530)
Co-authored-by: Kai-Tao Xie <kaitaoxie@qq.com>
* Hot fix rrp slow (#2526)
* review update
* updated the 2nd review comments
* String formatting
* fix multi_robot_test (#2524)
* Update README.md
* remove old build tools (#2532)
* Fix out of voxel grid array regression (#2460)
* Update dwb_local_planner.hpp (#2533)
Add remarks
* nav2_bringup reorder (#2527)
* nav2_bringup reorder
* change to trigger
* manual remove cmd
* revert
* add manual rm comd
* change
* change
* change
* change
* change
* change
* revert
* change
* change
* change v1 to v2
* try --cmake-clean-cache
* try --cmake-clean-cache
* v2 cache
* v3 cache
* Add new test for smac_planner_2d (#2531)
* Add new test for smac_planner_2d
* Fix costmap initialization
* Change set_parameters() with set_parameters_atomically()
* Improving coverage
* Remove debug helper method
* Fix linting issue introduced in https://github.com/ros-planning/navigation2/pull/2533 (#2539)
* [All 2D planners] Fix last pose orientation, fix small path and add ignore_goal_orientation parameter (#2488)
* end point orientation
* add ignore_goal_orientation to dyn param
* back() instead of [path.poses.size() - 1]
* Smac2d use_final_approach_orientation name
* add in NavFn
* add in theta_star
* wip
* deal with navfn double end pose case
* add tests
* back and revert test smac2d
* get path hotfix
* navfn wip
* Corner cases
* Checks consistency accross the 3 2D planners
* comment
* interpolate final orientation in planner instead of smoother
* clarify and homogenize comments
* using same cell test instead of distance compared to resolution
* remove unneeded else
* lint
* fix smac2d goal orientation override
* fix and add tests
* update comment
* typo
* simplify if (_use_final_approach_orientation) block
* Use worldToMapEnforceBounds in clear_costmap_service (#2544)
to avoid buffer overflow
Signed-off-by: zouyonghao <yonghaoz1994@gmail.com>
* Add new test for nav2_regulated_pure_pursuit (#2542)
* unit test for findDirChg
* lint fix
* add test for unchanged direction
* fix for loop ending
* Enabled runtime configuration of parameters for Hybrid A* (#2540)
* Enabled runtime configuration of parameters for Hybrid A*
* Fix parameter name
* Fix parameter type
* fix pf_ use-before-initial in laserReceived() callback (#2550)
Co-authored-by: Kai-Tao Xie <kaitaoxie@qq.com>
* add semantic checks (#2557)
Co-authored-by: Kai-Tao Xie <kaitaoxie@qq.com>
* Fix deprecated type conversion (#2538) (#2543)
Replace implicit deprecated type conversion from rclcpp::Client::FutureAndRequestId to std::shared_future with call to convenience method FutureAndRequestId::share()
Co-authored-by: Nikolay Panov <pn.interface@gmail.com>
* Register nodes as components (#2562)
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* Updates to Nav2 Theta Star Planner docs (#2559)
* - removing 'w_heuristic_cost_' from Nav2 Theta Star Planner as a user configurable parameter
- updates to `nav2_theta_star_planner/README.md` to remove its reference wherever applicable
- updates to `nav2_theta_star_planner/src/theta_star_planner.cpp` to set `w_heuristic_cost` as `std::min(1.0, w_euc_cost_)`
- fixed an incorrect check of whether the start and goal pose are the same in `nav2_theta_star_planner/src/theta_star_planner.cpp`
* Update theta_star_planner.hpp
clean up stuff from forced merge
* Update theta_star_planner.cpp
clean up stuff from a forced merge
* Update README.md
* Update theta_star_planner.cpp
Re-adding the goal to ensure, that the last pose, is the goal exactly
* fix linting issues
* Update theta_star_planner.cpp
Co-authored-by: Steve Macenski <stevenmacenski@gmail.com>
* Fixed vector::reserve exception in smac planner due to precision error (#2563)
- Related issue: https://github.com/ros-planning/navigation2/issues/2547
* Fixing minor build warnings (#2564)
* Fixed test_bt_navigator_with_groot_monitoring (#2568)
* removed redundant params in smac example (#2570)
* Support manual composed bringup for Nav2 applications (#2555)
* support manual composed bringup for Nav2 applications
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* fix linters
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* update
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* rm composed_bringup_launch.py
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* Update nav2_bringup/src/composed_bringup.cpp
* Update nav2_bringup/src/composed_bringup.cpp
* Update nav2_bringup/src/composed_bringup.cpp
* Update nav2_bringup/src/composed_bringup.cpp
* Update nav2_bringup/src/composed_bringup.cpp
* Update nav2_bringup/src/composed_bringup.cpp
* Update nav2_bringup/README.md
* Update nav2_bringup/README.md
* Update nav2_bringup/README.md
* Update nav2_bringup/README.md
* update nav2_bringup/README.md
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* remove unused use_lifecycle_mgr
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
Co-authored-by: Steve Macenski <stevenmacenski@gmail.com>
* Fixed model loading issue with Gazebo and Rviz2 (#2577)
1. Update uri(model://) of dae files for waffle model in waffle.model(Gazebo uses)
2. Update uri(package://) of dae files for turtlebot3_waffle.urdf(Rviz2
uses)
* [Draft] Set parameters result example (#2576)
* change dyn param method
* typo
* remove unnecessary log
* handler init in activate
* Fix typo (#2583)
fix typo in warning msg;
* The smac planner max_planning_time parameter can timeout the A* planning stage (#2586)
- Until now the max_planning_time parameter only had an effect on the smoothing and sampling stages of the planner, so the actual A* planning could possibly take a lot longer
* Change dynamic parameters method (#2585)
* dyn param method
* dont add to reset for checkers
* fix test
* Resolving 2453: Adding timeout for recoveries in case of stuck conditions (#2589)
* allowing backup and spin to have timeouts in case of error / getting stuck
* fix linting
* Removing multi-robot test, flaky in CI and unneeded (#2601)
* Create mergify.yml (#2602)
* Create .mergify.yml
* Rename .mergify.yml to mergify.yml
* Fixes for Windows into main (#2603)
* Fix nav2_util
* Don't include libgen.h on Windows
* Replace sleep with Sleep on Windows
* Fix rviz plugin
Undef NO_ERROR from windows headers
* Fix smac_planner
* Define _USE_MATH_DEFINES for M_PI
* Replace uint with unsigned int
* Don't add compiler flags that aren't supported on MSVC
* Use vector instead of VLA when building with MSVC
* Fix dirname for \ paths
* Fix typo
* Fix linter
* Delete head branch automatically on PR merge (#2608)
* Modified the projection_time calculation (#2609) (#2610)
(cherry picked from commit 8faa0f0550027edc7bf2640f23cff13b2af60a8f)
Co-authored-by: Narahari Rahul Malayanur <60045406+Naraharirahul@users.noreply.github.com>
* Update mergify.yml (#2612)
* Update mergify.yml
* Update mergify.yml
* Update mergify.yml
* Update mergify.yml
* Update mergify.yml
* Update mergify.yml (#2616)
* Update mergify.yml for Foxy backports and maintainer exceptions (#2620)
* Update mergify.yml
* Update mergify.yml
* adding dynamic params to RPP and cleaningup style in others (#2607)
* adding dynamic params to RPP and cleaningup style in others
* adding tests
* tests working
* adding alot more read me context and updates (#2625)
* Stopping networking-related BT failures from failing task, rather fail specific BT node (#2624)
* addig try/catch around non-tree exceptions to fail node
* Update bt_action_node.hpp
* linting
* Updating mergify rules for capitalization of usernames (#2630)
* adding groot separation by BT type (continuation on 2409) (#2627)
* adding groot separation by BT type
* fixing linting
* refactor of #2624 for simplicity (#2629)
* removing kinematic limiting from RPP (#2631)
* removing kinematic limiting from RPP
* update tests
* Resolving 2574: Smac 2D on map (#2588)
* Fixed model loading issue with Gazebo and Rviz2
1. Update uri(model://) of dae files for waffle model in waffle.model(Gazebo uses)
2. Update uri(package://) of dae files for turtlebot3_waffle.urdf(Rviz2
uses)
* fixing off map expansions
Co-authored-by: aliben <aliben.develop@gmail.com>
* Add dynamic parameters do Costmap2DROS (#2592)
* dyn params for costmap2dros
* simplify dyn footprint padding
* add test
* updated header for rolling
* Bug fix: set the BT Navigator "default_nav_to_pose_bt_xml" and "default_nav_through_pose_bt_xml" from the yaml file (#2664)
* removing the has_parameter() and replacing with the get_parameter()
* fixing navigate_through_poses as well
* Adding comment to clarify intent
* Refactored amcl code for modularization of motion models(main branch) (#2642)
* Delete nav2_amcl directory
* Add files via upload
* Remove "."
* Update CHANGELOG.rst
* Remove redundant code from CMakelists.txt
* Reverted amcl_node.cpp
* Cleaned up differential_motion_model.cpp
* Update plugins.xml
Removed redundant code,to make both the models under one library.
* Cleaned up omni_motion_model.cpp
* Cleaned up motion_model.cpp
* Delete CHANGELOG.rst
* Update CMakeLists.txt
* Update motion_model.hpp
* Add header files via upload
* Update amcl_node.cpp
* Update differential_motion_model.cpp
* Update omni_motion_model.cpp
* Update CMakeLists.txt
* Update differential_motion_model.hpp
* Update motion_model.hpp
* Update omni_motion_model.hpp
* Update amcl_node.cpp
* Update differential_motion_model.cpp
* Update motion_model.cpp
* Update omni_motion_model.cpp
* Update differential_motion_model.cpp
* Update differential_motion_model.hpp
* Update motion_model.hpp
* Update omni_motion_model.hpp
* Update differential_motion_model.cpp
* Update motion_model.cpp
* Update omni_motion_model.cpp
* Update differential_motion_model.hpp
* Update motion_model.hpp
* Update omni_motion_model.hpp
* Update motion_model.cpp
* Update motion_model.cpp
* Update nav2_params.yaml
* Update CMakeLists.txt
* Update nav2_multirobot_params_2.yaml
* Update nav2_multirobot_params_1.yaml
* Update speed_global_params.yaml
* Update speed_local_params.yaml
* Update keepout_params.yaml
* Footprint collision checking for nav2_regulated_pure_pursuit (#2556)
* initial
* initial
* initialize
* un
* working
* lint fix
* wrk
* clean_up
* remove code repettion
* rm unused variable
* initialize footprint on configure itself
* revert to defaulted constructer
* remove rviz change
* remove unneccessary changes
* review changes
* define traverse unkown
* set true
* reviw changes
* styling changes
* collision checking condition
* used back()
* make loop end condition fixed
* review changes
* fix
* removed midpose curvature
* removed stop collision check at goal pose
* Windows fix (#2672)
* Add support for reconfiguring params using GUI in ThetaStarPlanner (#2665)
* - removing 'w_heuristic_cost_' from Nav2 Theta Star Planner as a user configurable parameter
- updates to `nav2_theta_star_planner/README.md` to remove its reference wherever applicable
- updates to `nav2_theta_star_planner/src/theta_star_planner.cpp` to set `w_heuristic_cost` as `std::min(1.0, w_euc_cost_)`
- fixed an incorrect check of whether the start and goal pose are the same in `nav2_theta_star_planner/src/theta_star_planner.cpp`
* Update theta_star_planner.hpp
clean up stuff from forced merge
* Update theta_star_planner.cpp
clean up stuff from a forced merge
* Update README.md
* Update theta_star_planner.cpp
Re-adding the goal to ensure, that the last pose, is the goal exactly
* fix linting issues
* Update theta_star_planner.cpp
* update the README and code to prevent errors
* add support for reconfiguring params using GUI
* Update theta_star_planner.hpp
* fix linting issues
* remove being set as the param
* Update theta_star_planner.cpp
* Update theta_star_planner.cpp
* Update theta_star_planner.cpp
* Update theta_star_planner.cpp
* Update theta_star_planner.cpp
Co-authored-by: Steve Macenski <stevenmacenski@gmail.com>
* Corrected reinitialization in nav2_smac_planner::HybridMotionTable (#2680)
- The HybridMotionTable did not reinitialize properly in case the min_turning_radius or the motion_model was changed.
* State Lattice Planner Work Revival (#2582)
* adding state lattice beginning point
* refactoring to remove analytic expansion into its own object
* moving smoother and collision checker into cpp files
* allowing smac planner to work with irregular bin distributions
* integrating @jwallice42 s work and enabling reverse expansions
* adding in intermediate collision checking and backtracing
* adding traversal function
* adding test coverage and passing for node lattice
* adding updated rosdep key
* compiling, generating paths (though bogus)
* adding working base code
* adding search info
* fully working for non-lateral lattices
* removing downsampler for lattice, pushing in cost penalty
* adding in updated heuristics, tuned gains, and generally working for ackermann cases
* reverse expansion working
* adding complete support, refactoring, etc
* adding forgotten files
* adding readme updates
* adding max planning time to state lattice
* adding dynamic parameters
* removing extra config file
* fix linting
* updating test
* adding several unit tests
* adding system test
* fixing crash
* adding default model
* adding corrected exporting of plugins
* adding more test coverage
* reorder tests to see if natters
* unlink motions_lib plugin library in nav2_amcl (#2675)
* separate amcl_core library from motions_lib plugin library
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* update plugins.xml
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* update default value of robot_model_type
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* debugging flaky shutdown on plugin planner tests (#2682)
* debugging flaky shutdown on plugin planner tests
* adding independent tests
* adding types
* adding new QoS and fix param declares
* try this
* try tihis
* try 10hz
* try sleeping on setup
* add footprint pub
* adding additional pub for debug
* try adding a spin for debug types
* Fix spin recovery test (#2684)
* debugging flaky shutdown on plugin planner tests
* adding independent tests
* adding types
* adding new QoS and fix param declares
* try this
* try tihis
* try 10hz
* try sleeping on setup
* add footprint pub
* adding additional pub for debug
* try adding a spin for debug types
* replacing spin with sleep
* Updating regex for mergify (#2691)
* Smoother server (#2569)
* old version of smoother server with path truncation and without collision checker and footprint subscriber
* Removed path truncation from smoother server, added footprint subscriber
and path collision checking
* smoother code styling fixes
* Update nav2_tree_nodes.xml
* WIP: PR #2569 fixes
* smoother server changes as requested in PR #2569, tests added
* cpplint fixes
* added nav2_smoother to composed bringup
* smoother server: improved test coverage, details fixed
* removed smoother from composed bringup lifecycle nodes as it fails on
non-existing plugins
* removed smoother server default plugin since it's not merged yet, readded into bringup
* nav2_smoother tests: added mocked class loader to improve
nav2_smoother.cpp test coverage
* nav2_smoother 100% test coverage
* test_smoother_server cpplint fix
* test_smoother_server added cleanup transition
* nav2_smoother added default plugin type autoconfig test
* Update nav2_smoother/README.md
* removed smoother_server_rclcpp_node from nav2_params.yaml, added
smoother_server to other config YAML files
Co-authored-by: afrixs <matej.vargovcik@gmail.com>
Co-authored-by: Steve Macenski <stevenmacenski@gmail.com>
Co-authored-by: Matej Vargovcik <vargovcik@robotechvision.com>
* removing dead code from AMCL (#2693)
* Lifecycle startup fix (#2694)
* removed init_timer from lifecycle callback group
* lifecycle manager: startup() on autostart_ calling from callback_group_
(Bond::waitUntilFormed needs node spinning in the background)
Co-authored-by: Matej Vargovcik <vargovcik@robotechvision.com>
* If not downsampling costmap, default factor to 1 (#2695)
* removing live groot monitoring from Nav2 (#2696)
* removing live groot monitoring from Nav2
* rename test
* fixing linting issues
* readd default
* linting
* Remove "wrong"/unnecessary dependency (#2700)
* Theta star planner fix (#2703)
* add test to check if the goal is unsafe with theta star planner
Signed-off-by: Daisuke Sato <daisukes@cmu.edu>
* check if a plan is generated (#2698)
Signed-off-by: Daisuke Sato <daisukes@cmu.edu>
* adding max analytic expansion length (#2697)
* adding max analytic expansion length
* fix tests and add dynamic parameter support
* [RPP] remove velocity scaling option (now required) and enable orientation sp projection times (#2701)
* remove velocity scaling option (now required) and enable orientation sp projection times
* get max radius
* Version bump up colcon extensions (#2706)
* Always initialize clearing endpoints in voxel layer (#2705)
* Prohibit of simultaneous usage of callbacks accessing to pf_ object (#2712)
* CostmapTopicCollisionChecker tf stamp fix (#2687)
* fixed tf lookup in costmap_topic_collision_checker unorientFootprint method
* minor cleanup
* uncrustified
* fixed stamps for tests which include CostmapTopicCollisionChecker
* updated CostmapTopicCollisionChecker construction for nav2_smoother
* cleaned up CostmapTopicCollisionChecker and FootprintSubscriber
* global_frame declaration returned to recovery_server as it is needed by
recovery plugins
* collision checker params updated
* Made transformFootprint same-input-output-safe
* uncrustified footprint.cpp
* CostmapTopicCollisionChecker enabled to fetch footprint only once per
multiple checks (same like costmap)
* CostmapTopicCollisionChecker: updated doxygen
* Update nav2_costmap_2d/include/nav2_costmap_2d/costmap_topic_collision_checker.hpp
Co-authored-by: Steve Macenski <stevenmacenski@gmail.com>
* Update costmap_topic_collision_checker.hpp
Co-authored-by: afrixs <matej.vargovcik@gmail.com>
Co-authored-by: Matej Vargovcik <vargovcik@robotechvision.com>
Co-authored-by: Steve Macenski <stevenmacenski@gmail.com>
* adding dynamic parameters support for controller, planner, and waypoint servers (#2704)
* adding dynamic parameters support for controller, planner, and waypoint servers
* reset dynaic params on deactivate
* adding mutex to dynamic parameters support for 3 servers
* removing waypoint follower lock
* kill composition for 1 test
* adding test coverage for dynamic parameters in main task servers
* changing header for composed bringup
* adding backwards state update (#2717)
* Clean .gcda files for invalidated package tests (#2715)
* Clean .gcda files for invalidated package tests
"Each time you run the program, the execution counts are summed into any existing .gcda files, so be sure to remove any old files if you do not want their contents to be included."
https://llvm.org/docs/CommandGuide/llvm-cov.html
* Clean .gcno files for invalidated package builds
"At the time you compile the instrumented code, a .gcno data file will be generated for each object file."
https://llvm.org/docs/CommandGuide/llvm-cov.html
* Bump up cache for testing
* Revert .gcno cleaning
cmake should handle stale files
* Explicitly test uppassed tests as well
keeping TEST_FAILURES even if redudnet,
as it's still useful to echo for diagnostics
* Bump up cache for testing
* Cache test results
to avoid this error with colcon test-results:
Path 'test_results' does not exist
* Bump up cache for testing
* changing formulation of Hybrid-A* to be admissible and retune (#2713)
* adding shim rotation controller plugin (#2718)
* adding shim rotation controller plugin
* adding to metapackage
* adding max kinematic accel
* removing old note
* adding video to readme
* adding collision detector
* adding projection time for collision checking to rotation shim controller
* add another test case
* adding dynamic param tests for sim ahead time
* Update PULL_REQUEST_TEMPLATE.md
* Update mergify.yml
* Smac anomalies fix (#2724)
* fixed analytic expansion poses skipping and restoring of original pose
if neighbor not interesting
* Revert "fixed restoring of original pose"
This reverts commit a4f51ee28b03bf49f2cb63debfed49f7390fe6ef except of
analytic expansion skipping
* --
* removed API change, doing cleanup inside AnalyticExpansion
* smac analytic expansion: removed node skipping
* Update analytic_expansion.cpp
* fixed linting errors
* changed raw pointers to unique_ptrs
* Update analytic_expansion.cpp
* linter errors fix
Co-authored-by: Matej Vargovcik <vargovcik@robotechvision.com>
* Fixed min lookahead dist when moving backwards (#2726)
Co-authored-by: seasony-vp <72447461+seasony-vp@users.noreply.github.com>
* adding reversing to smoothers (#2727)
* adding reversing to smoothers
* smoother working
* fixing CI
* update test
* Smoother feasbility (#2731)
* first go at a new smoother feasibility
* adding doxygen
* adding fix for reeds shepp mini expansion issue
* removing comments
* removing old line
* reordering declaration
* fixing last bug
* phew, that was a hard bug to fix...
* adding test coverage
* increase general test coverage (#2735)
* increase test coverage
* adding more coverage
* adding more coverage
* Apt upgrade with new pkgs (#2737)
* Apt upgrade with new pkgs
using --with-new-pkgs flag
to ensure all packages in CI image are updated
https://unix.stackexchange.com/a/241062/213124
* Pipe stdout from apt upgrade to log file
for better debuging and transparency
* adding dynamic parameter tests to navfn (#2743)
* RPP max lookahead collision detection time limited by lookahead point itself (#2741)
* adding RPP max lookahead collision detection time limited by lookahead point itself
* linting
* linting
* allowing rviz to visualize laser scans again with rolling binaries (#2745)
* increasing backup distance and speed (#2746)
* increasing backup distance and speed
* increasing speed
* Update colcon and install lld (#2761)
* Use LLD linker (#2734)
* Use LLD linker
https://lld.llvm.org
https://github.com/colcon/colcon-mixin-repository/pull/28
* Bump cache for testing
* docs: clarify that container images are for building, not deploying (#2763)
Signed-off-by: Adam Aposhian <aposhian.dev@gmail.com>
* Simplify coverage jobs for CI (#2762)
* Test coverage from release build
* Bump up cache for testing
* Merge test jobs
and name coverage uploads
using set RMW
* fix invert logic (#2772)
* Updated primitives file to latest version (#2774)
* Updated primitives file to latest version
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed tests for parsing and path checking
* Fix : misspelling in smac_planner's readme (#2778)
* Linting fix (#2779)
* Bugfix tf lookup of goal pose in nav2_controller (#2780)
Signed-off-by: Erik Örjehag <erik@orjehag.se>
* adding support for in-place rotations in state lattice (#2747)
* [costmap2d] make robot_base_frame param dynamic (#2782)
* make costmap2d robot_base_frame_param dynamic
* add test
* add rejecting logic and test
* support dynamic composed bringup (#2750)
* support dynamic composed bringup
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* update
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* update delcares in launch
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* use component_container_isolated
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* update README
Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
* State Lattice Planner Control Set Generation (#2587)
* Initial version of motion primitive generation using sampling
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Changed to lattice motion primitive generation
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* minor clarity fix for yaw equation
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* WIP - change of parameters + altering of step_distance
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Implemented simplified method using arcs and geometry
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added testing for trajectory generator
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed several bugs and changed trajectory generator to use step distance
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added left turn parameter
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Updated tester to match latest trajectory generator
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Implemented lattice generator
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added edge case for parallel lines at 90 degrees
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added generation of paths in all quadrants
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Update to parameter names and return values
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Bug fixes
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added sample config file
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added script for generating and writing to file
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed end pose in list not being actual end point
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Output now sorted by startAngle then endAngle
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Start level calculated by flooring instead of rounding
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Tidy code and added indentation to output file
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed start/end angle being set to 180 or -0
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added dataclasses instead of passing around tuples
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added yaws to path
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Updated tests
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed bug in line path and added function comments
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Updated to reflect changes in trajectory generator
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed missing negative sign for start angles
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed min trajectory length being calculated wrongly
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed negative zero appearing in poses
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Set step distance to always be 10% of grid separation
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Explicitly create yaws as float64 type
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Updated tests to reflect step distance change
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added motion model handling
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Changed dataclasses to be frozen
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* step distance must now be passed in to trajectory generator
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added trajectory generation for all quadrants
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added logging
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added arc length and total length computed properties to trajectoryParams
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed trajectory generation for all quadrants
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added a method to create params for trajectories with no arcs
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Simplified lattice generation
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Moved some functions to a separate file
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Changed to use purely radians
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed -0.0 appearing again
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed angle difference calculation to account for turning direction
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Yaws properly generate when crossing -180/180 boundary
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Changed input/output parameters to be snake case
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Moved yaw interpolation to helper
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed rounding in outputs
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed paths to be relative to file location
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added a trajectory visualizer for easier testing/debugging
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Sorted primitives by angle starting at 0.0 and running CCW
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed path generation when distance was very small
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added R-Tree to speedup minimal path search
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added printing of time taken
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Changed max length to stopping threshold
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Use numpy functions instead of python functions for converting trajectories to list
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added a requirements.txt file
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed saving of visualizations to create the folder if it does not exist
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Removed unused import
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Generated a few samples for use
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Removed spaces from folder names
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed interpolating of yaws to use vectorization
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed bug causing differing precision in angles when flipping
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Plots now have same axis across trajectories
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Replaced function with lambda
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added a README
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Updated samples
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Refactored to follow pycodestyle
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Commented code and adjusted to match pycodestyle
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Commented code and adjusted to match pycodestyle
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Moved MotionModel class into LatticeGenerator
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Changed start/end angle in output to be index of angle in heading_angles
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Sample primitives now only contain 5cm x [0.5m, 1m] x [ackermann, diff,
omni]
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added config file parsing
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Updated README to include file structure details
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Renamed level to wavefront for better clarity
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added Apache 2.0 License
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Changed how paths are passed in and updated the output file
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Updated sample primitives and added visualizations
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added rotation matrix helper function
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added additional helper functions
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed path spacing issues
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Changed member variables of Parameter dataclass
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Changed how path spacing is handled
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Updated docstrings and type annotations
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added printing of distances between points for trajectory visualizer
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed bug in omni and diff model generation
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Regenerated all sample files using latest version
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added more comments for better clarity
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Updated readme for better clarity
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added recommended values to parameters
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Refactored TrajectoryPath and added docstrings for classes
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added comments for purpose of script
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Added docstring to classes
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Removed leftover print statement
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Moved version constant to separate file
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Fixed linting errors
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Removed backslashes in docstrings and made summaries one-liners
Signed-off-by: Matthew Booker <mattbooker97@gmail.com>
* Smac planner admissible obstacle heuristic (#2765)
* added obstacle_heuristic_admissible, change_reverse_penalty,
max_analytic_expansion_angle_range,
max_analytic_expansion_cost_subelevation and max_analytic_expansion_length options
* a_star test build fixed
* tests partially fixed
* --
* tests fixed
* obstacle heuristic improved
* removed max_analytic_expansion_length
* returned heuristic costmap downsampling
* cleanup before feature separation
* change_reverse_penalty feature separation
* analytic_expansion_constraints feature separation
* smac smoother bypass separation
* smac planner debug separation
* uncrustified smac planner changes
* improved smac admissible heuristic code readability
* fixed cleanup errors
* smac planner: added retrospective_penalty for speedup
* smac retrospective penalty: fixed errors, adjusted value in readme
* fixed doxygen, added new params to tests
* fixed smac tests
* Update node_hybrid.cpp
* removed inadmissible heuristic computation, updated names + signedness
to match previous implementation
* Smac: updated README.md
* Smac planner: setting retrospective_penalty default to 0.025, updating
readme and costmap_downsampler code styling
* Smac: fixed out-of-map cell obstacle heuristic crash, updated tests
* Smac: fixed merge errors and test values
* Smac: reorganized obstacle heuristic method code
* Smac: removed redundant check
* Update nav2_smac_planner/src/smac_planner_hybrid.cpp
Co-authored-by: afrixs <matej.vargovcik@gmail.com>
Co-authored-by: Matej Vargovcik <vargovcik@robotechvision.com>
Co-authored-by: Steve Macenski <stevenmacenski@gmail.com>
* Updating smoother for unknown space (#2791)
* adding an abstract BTNode for action cancellation (#2787)
* adding an abstract BT for action cancellation
* review update and adding test
* adding controller cancel BT Node
* Fixing the copyrights
* review updates
* satisfying lint
* adding the bt to the defaults
* update
* Cleaned up version of is path valid (#2591)
* Working replanning if path is vaild, however there is a bug with new goal points
* Completed tested of path replanning
* Delete uncrustify.cfg
* Clean up for code review
* steady
* small code review changes
* listen to server_timeout in input port
* Lint issues
* reverted file for navigate to pose
* lint fixes
* added global update condition
* headless
* revert default bt
* Add image of new bt tree
* circle ci and comments
* add test for globally updated goal condition
* added plath replanning only if path is invalid to navigate to pose
* Added nodes to nav2_tree_nodes
* cleanup BT tree for invalid path
* code fixeds
* lint fixes
* added default timeout
* lint fix
* server timeout fix
* removed images of new BTs
* added rate controller
* lint fix
* Moved server timeout to constructor
* fixed xml lint error
* fixed test for is path valid
* new bt image for reference
* remove reference
* Feature to call controller action server to follow path (#2789)
* Added call to controller action server with a path to follow
++ Added a new function in robot_navigator,py
++ Added a launch script to test function
++ Updated setup.py to include demo_follow_path.py
* Code refactoring
* Code refactoring
* Code refactoring for consistency
* Updated README.md
* Resolved executable conflict in setup.py for example_follow_path.py
* Code refactoring with ament_flake8
* BT nodes xml naming fix (#2805)
* Add recoveries to simple commander (#2792)
* fix Typo
* add recoveries
* add docs
* added demo using backup and spin
* rename isNavigationComplete to isTaskComplete
* rename cancelNav to cancelTask
* add prints
* fix premature exit
* rename NavResult to TaskResult
* fix readme order
* fix import order
* renaming
* renaming
* added planner_id (#2806)
* Fixing the issue #2781: raytraceLine with same start and end point (#2784)
* Fixing the issue #2781: raytraceLine with same start and end point no longer causes segmentation fault
* Some whitespace modifications to make the code pass release_test
* Add testcase for raytraceLine the same point
Co-authored-by: Alexey Merzlyakov <60094858+AlexeyMerzlyakov@users.noreply.github.com>
* Add optional node names to wait (#2811)
* add option to specify navigator and localizer to wait for
* add docs for waituntilNav2Active
* wait for pose only for amcl
* revert order
* Drop unneeded dependency on python-enum34 (#2819)
The enum module was added in PEP 435 and is present in Python versions
3.4 and newer, which includes all currently supported Python versions.
* remove resizing on update message callback (#2818)
* smoother fix (#2822)
* Dynamic parameter setting of costmap layers (inflation/static/voxel) (#2760)
* added dynamic params to inflation layer
* Fixed codestyle divergence
* Added test for inflation param setting
* Added dynamic parameter setting to voxel layer
* Fixed code styling
* Added test for parameter setting of voxel layer
* Added dynamic parameter setting for static layer
* Added test for dynamic param setting of static layer
* turtlesim3 simulation testing bugfixes
* Imported some after param change function calls from ROS1 for all the layers
* Updated changes to the main branch
* Review changes by Alexey
* Review comments by Alexey 2
* Fixed codestyle divergence
* Removed blank line
* Review comments by Alexey 3
* Removed blank lineS
* Review change by Steven
* restrict test_msgs to test_dependency (#2827)
Signed-off-by: Alberto Soragna <alberto.soragna@gmail.com>
* remove unused odometry smoother in bt navigator (#2829)
* remove unused odometry smoother in bt navigator
Signed-off-by: Alberto Soragna <alberto.soragna@gmail.com>
* reorganize bt navigator to shared odom_smoother object with servers
Signed-off-by: Alberto Soragna <alberto.soragna@gmail.com>
* fix: modify build failure of nav2 bt navigator (#2836)
Signed-off-by: wep21 <border_goldenmarket@yahoo.co.jp>
* Add all action options (#2834)
* added options
* update docs
* edge fix (#2840)
* edge fix
* code review
* Launch single (#2837)
* found error
* complete test setup
* updated metrics
* reverted planner server
* revert launch changes
* remove edits
* reverted nav2_bringup
* revet robot navigator
* add spaces
* removed map
* linting fix
* removed package, moved valuable stuff to tools
* addede more planners
* planner benchmarking
* adding stable dynamic reconfigure behavior (#2843)
* Update RPP readme
* Adding theta* to the main packages list
* Ci focal rolling (#2851)
* use rolling-ros-base-focal
* hotfix: remove rosdep update from ci
* Fix large angle changes in lattice primitive generator (#2831)
* Changed normalize_angle to exclude -pi and include pi
Signed-off-by: mattbooker <mattbooker97@gmail.com>
* Run normalize_angle on yaws to fix angle changes
Signed-off-by: mattbooker <mattbooker97@gmail.com>
* Added testing for the Lattice Generator
Signed-off-by: mattbooker <mattbooker97@gmail.com>
* Fixed linting errors
Signed-off-by: mattbooker <mattbooker97@gmail.com>
* Changed normalize_angle to exclude -pi and include pi
Signed-off-by: mattbooker <mattbooker97@gmail.com>
* Run normalize_angle on yaws to fix angle changes
Signed-off-by: mattbooker <mattbooker97@gmail.com>
* Added testing for the Lattice Generator
Signed-off-by: mattbooker <mattbooker97@gmail.com>
* Fixed linting errors
Signed-off-by: mattbooker <mattbooker97@gmail.com>
* Updated sample primitive files
Signed-off-by: mattbooker <mattbooker97@gmail.com>
* Removed default_model.json file (duplicate of sample primitives)
Signed-off-by: mattbooker <mattbooker97@gmail.com>
* Fixed issue caused by merge
Signed-off-by: mattbooker <mattbooker97@gmail.com>
* Replaced default primitive files
Signed-off-by: mattbooker <mattbooker97@gmail.com>
* Removed whitespace in folder names
Signed-off-by: mattbooker <mattbooker97@gmail.com>
* Moved angle normalization to occur before output instead of during generation
Signed-off-by: mattbooker <mattbooker97@gmail.com>
* Adjusted angle normalization to output in the range of [0, 2pi)
Signed-off-by: mattbooker <mattbooker97@gmail.com>
* Updated primitive files with latest changes
Signed-off-by: mattbooker <mattbooker97@gmail.com>
* Fixed normalize_angle docstring
Signed-off-by: mattbooker <mattbooker97@gmail.com>
* Fix: bt_navigator crashes on lc transitions (#2848)
* fix empty address access on halt all actions
* fix unsafe declaration of parameters
* restore odom smoother
* fix styling issues
* add missing semicolumn
* fix-collision checker must capture lethal before unknow (#2854)
* Bound closest pose by costmap and turnaround point (#2812)
* created integrated distance util
* chng transform begin , end and always limit lp pt
* chg trnsfm begin end and use euclidean distance
* linting fix
* linting fix
* limit trnsfm to cusp
* remove print
* linting fix
* lint fix
* hypot
* use direction change as upper bound for closest pose
* bound closest waypoint by direction change
* use max_transform_dist for closest_pose_upper_bound
* remove first_element_beyond optimization todo
* added path_utils for test path generation
* initial test for path_utils
* added full failing test for path_utils
* fixed path_utils segfault
* properly initialize straight transform
* fixed right turn infinite loop
* fixed path_utils transforms for half-turns
* all path_utils tests passing
* added half turn test for path_utils
* remove unused dependencies from test_path_utils
* move path_utils classes into path_building_blocks namespace
* use hypot for euclidean_distance
* rename findDirectionChange to findCusp
* use integrated distance for lookahead distance
* Revert "use integrated distance for lookahead distance"
This reverts commit 6e4c6cd3a795faf29f977df7316b5d58d9c1f135.
* parameterize transformation_begin bound
* change default to a regular constant
* use std::hypot for x, y, z
* initial failing test for transformGlobalPlan
* refactor transform global plan into separate test fixture
* make TransformGlobalPlan test fixture, and fix transforms
* no_pruning_on_large_costmap test passing
* added another transformGlobalPlan test
* added a test for all poses outside of costmap
* added good shortcut test for transformGlobalPlan
* added more tests for rpp costmap pruning
* cpplint
* remove unused rclcpp::init and rclcpp::shutdown
* change default max_distance_between_iterations to max_costmap_extent
* rename max_distance_between_iterations to max_robot_pose_search_dist
* increase default dwb prune distance
* add initial docs for max_robot_pose_search_dist to README
* add note about when to set max_robot_pose_search_dist
* rename first_element_beyond to first_after_integrated_distance
* renamed findCusp to findVelocitySignChange
* move path_utils to RPP tests
* only check velocity sign change when reversing is enabled
* do not check cusp for dwb transformed plan end
Co-authored-by: sathak93 <sathak0730@gmail.com>
Co-authored-by: Adam Aposhian <adam.aposhian@fireflyautomatix.com>
* Apply raytraceLine 3D fixes to its 2D version (#2857)
(back-port of #2460 and #2784)
* adding Bt to cancel wait, spin and backup (#2856)
* adding Bt to cancel wait
* Adding cancel for spining recovery
* adding BT for cancelling the backup behavior
* adding cancel recoveries to param files for wider test coverage
* updating BT's XML
* updating XML nodes
* Fix cpplint include order errors (#2860)
* fix nav2_smac_planner cpplint
* fix dwb_core cpplint
* fix nav2_map_server cpplint
* fix nav2_waypoint_follower include order
* fix nav_2d_utils include order
* Linter fixes (#2861)
* satisfying uncrustify
* cpplint fixes
* Additional revision for costmap_bresenham_2d test (#2862)
* Removing old unused function and comment (#2863)
* fix core linting errors (#2864)
* fix core linting
* fix smac tests
* fix filepath
* fix angle normalization
* fix smac testes
* finish lint
* euclidean distance: by default use 2d mode (#2865)
Signed-off-by: Vinny Ruia <vinny.ruia@fireflyautomatix.com>
* add the missing ports and description for the XML nodes (#2866)
* add the missing ports and description
* update
* update
* Fixed first_map_only parameter issue (#2870)
* Fixed first_map_only parameter issue
Signed-off-by: Harshal Deshpande <hardesh1deshpande@gmail.com>
* Adressed comments on PR
Signed-off-by: Harshal Deshpande <hardesh1deshpande@gmail.com>
* Fixed linting error
Signed-off-by: Harshal Deshpande <hardesh1deshpande@gmail.com>
* Mixed replan (#2841)
* working global time expired node
* test for global_time_expired_condition
* revert navigate to pose
* added '
* added space
* fix
* code review
* added 15 sec replanning time
* new bt
* revert bt
* fix
* reset timer when path is updated
* name change and fixes, it works!
* completed tests
* revert
* code review
* code review
* replanning time adjustment, rename
* removed commented code
* test fix
* remove comment
* Adding a new BT node Is_path_longer_on_approach (#2802)
* adding behavior and updating BT tree
* update BT
* updating copyrights
* fixing wrong test fixture
* review update
* update comments
* review update
* review update after testing
* Update nav…
7 tasks
savalena
pushed a commit
to savalena/navigation2
that referenced
this pull request
Jul 5, 2024
* support dynamic composed bringup Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com> * update Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com> * update delcares in launch Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com> * use component_container_isolated Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com> * update README Signed-off-by: zhenpeng ge <zhenpeng.ge@qq.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Signed-off-by: zhenpeng ge zhenpeng.ge@qq.com
Basic Info
Description of contribution in a few bullet points
use dynamic composition instead of manual composition which exists in Nav2 currently.
SmootherServernode as component.containerinbringup_launch.py.LoadComposableNodesandComposableNodeto load Nav2 node inlocalization_launch.pyandnavigation_launch.py.performance test
component_container_mt)dynamic composition consumes lower memory(saves ~70%), but consumes the similar cpu as normal multiple processes, due to the bad performance of
MultiThreadedExecutor.Future work that may be required in bullet points
component_container_isolatedinstead ofcomponent_container_mt, the feature (support dedicated executor for component nodes ros2/rclcpp#1781) was merged, but it isn't available now, we can use this method after the next rolling sync.For Maintainers: