* make the 3 params changeable at runtime

* use parameter events callbacks

* doxygen

* lint

* Fix recovery action collision check.

* Fix linting issue.

* Adding RPP to metapackage.xml

* Remove excessive exec depend

* Write doxygen for navfn

* Remove forward slashes

(WIP) Recoveries always return success, regardless if they're able to complete their required tasks or not (#1855)

* modified such that if the recovery is aborted due to potential collision, the return status will be FAILED

* Changed from SequenceStar to RoundRobin to mask any failure that might occur in the execution of the recovery action

* Change Backup recovery test such that the aborted recovery is expected and checked for

* fixing linting error

* Change depricated argument for backup recovery test

* added backup to the recovery actions, using defaults from the BackUp header

* Update nav2_bt_navigator/behavior_trees/navigate_w_replanning_and_recovery.xml

Co-authored-by: Steve Macenski <stevenmacenski@gmail.com>

* Reverting changes to BT. Seq. Star control node is in line to what we want

* checkpoitn commit

* Changed collision failure critera as backup collision abortion occurs @ -0.2 m as well. This was previously not caught as  the server always returned SUCCESS

* bool to control whether fake costmap should be generated

* added fake tester nodes

* Fake costmap passed through environmental variable now

* Fake Footprint being sent out

* breadcrumbs of fake spin tests

* fake odom

* launch file was messed up

* still troubleshooting spin

* trying to fake spin, goal keeps on getting rejected

* goal is now being accepted

* Fake spin not working well when asking for PI radians out

* playing around with timing wehn publishing fake transforms

* BT changes, testing TBD

* [WIP] Fix CI build issues (#2076)

* update to use on my fork to test changes to bondcpp

* using chrono literals in lifecycle manager

* nav2_rviz_plugins using chrono literals for API change

* using chrono literals in costmap_2d package from API change

* using chrono literals in observation buffer

* chrono literals for tests costmap

* chrono literal API changes

* changing API

* changing API

* API changes

* API change

* API change

* API change

* API change test

* API change test

* remove

* api updates

* update test values

* Patch for PhotoAtWaypoint plugin (#2067)

* nav2_way_point_follower; introduce photo at waypoint arrivals plugin

Signed-off-by: jediofgever <fetulahatas1@gmail.com>

* resolve cmake lint errors

Signed-off-by: jediofgever <fetulahatas1@gmail.com>

* resolve requested chages of first review

Signed-off-by: jediofgever <fetulahatas1@gmail.com>

* minor corrections on photo_at_waypoint header

* resolve requested changes of second review

Signed-off-by: jediofgever <fetulahatas1@gmail.com>

* update default save_dir

Signed-off-by: jediofgever <fetulahatas1@gmail.com>

* move directory checking code to initialize()

* add try catch block to catch possible execeptions while creating a directory for photo at waypoint plugin

Signed-off-by: jediofgever <fetulahatas1@gmail.com>

* adding logging and disabling

* Update nav2_waypoint_follower/plugins/photo_at_waypoint.cpp

Co-authored-by: Fetullah Atas <fetulahatas1@gmail.com>

* Update photo_at_waypoint.cpp

Co-authored-by: jediofgever <fetulahatas1@gmail.com>

* Fix for double free issue in map server testcases (#2078)

* Support in keepout filter for mask and costmap published in different… (#2054)

* Support in keepout filter for mask and costmap published in different frames

This fixes incorrect keepouts position issue when filter mask and current
costmap layer are published in different frames. This might appear
(but not restricted only to) when keepout filter is enabled for local costmap
with rolling window.

* Add transform initialization

* Enhance transform failure message

* Loop fix (#2068)

* Abort analytic expansion if crossing through already visited node

* Check that we are not creating an infinite loop at the goal node

* Mark nodes in analytic expansion as visited for the sake of completeness

* Move checking of already visited nodes to final stage of analytic expansion

* test if action server failures correctly propagate into BT context

* change recovery subtree so that RoundRobin replaces SequenceStar for good

* uncrustify and line nav2_behavior_tree changes

* uncrustify and linting

* adding clearing actions to own subtree

* more linting

* reduced testing value to see if that will stop the robot from aborting due to potential collision

* tighten up tolerance on backup since the requested backup is so small

* delint

* cleaned up spin recovery such that we simulate the robot slowly spinning into place, now I just have to simulate collision scenarios

* linting

* change order of Spin and Wait to match original

* fake spin failing due to potential collision even when empty costmap

* lint

* more lint

* can get all the spins to pass now

* spin test strangely passing all cases even though costmap is populated

* spin recovery

* commmit before I try to visualize this

* now I know that there are two things publishing to the same footprint

* 1) figured out conflicting pubs to footprint, 2) why can't I get costmap occupancy to cause collision

* Fake Spin Test shows failure correctly for angles greater than pi / 2. change launch file such that two publishers are not generating costmaps

* fake launch lint

* check in with linting to show the structure

* Add overall behavior tree system tests

Signed-off-by: Sarthak Mittal <sarthakmittal2608@gmail.com>

* Add RecoverySubtreeGoalUpdated BT test

Signed-off-by: Sarthak Mittal <sarthakmittal2608@gmail.com>

* Update readme and docs

* Add tests to check if BT XML files are well-formed

Signed-off-by: Sarthak Mittal <sarthakmittal2608@gmail.com>

* Uncomment tests that were commented out

Signed-off-by: Sarthak Mittal <sarthakmittal2608@gmail.com>

Co-authored-by: Steve Macenski <stevenmacenski@gmail.com>
Co-authored-by: jediofgever <fetulahatas1@gmail.com>
Co-authored-by: Alexey Merzlyakov <60094858+AlexeyMerzlyakov@users.noreply.github.com>
Co-authored-by: James Ward <james.ward@appliedev.com>
Co-authored-by: Sarthak Mittal <sarthakmittal2608@gmail.com>

* regulated pure pursuit migration commit

* adding speed limit API

* adding review comments + adding rotate to goal heading

* adding test dir

* add some initial tests

* more tests

* remove old comment

* improve readme

* fix CI

* first attempt at changing algos in tests

* allowing full path parameter substitutions

* adding integration tests

* enable SMAC testing too with new changes

* swap algos

* revert

* Update angular velocity after constraining linear velocity (#2165)

This ensures the robot moves towards the lookahead point more closely.
If the angular velocity is not updated, then the robot tries to take cuts while turning,
which could lead to collisions when near obstacles

Signed-off-by: Shrijit Singh <shrijitsingh99@gmail.com>

* Update cost scaling heuristic to vary speed linearly with distance (#2164)

* Update cost scaling to vary linearly with distance instead of relying on costmap cost

Signed-off-by: Shrijit Singh <shrijitsingh99@gmail.com>

* Resolve suggested changes

Signed-off-by: Shrijit Singh <shrijitsingh99@gmail.com>

* Add documentation for cost scaling parameters

Signed-off-by: Shrijit Singh <shrijitsingh99@gmail.com>

* Improve parameter descriptions

Signed-off-by: Shrijit Singh <shrijitsingh99@gmail.com>

* Comment cost scaling tests since layered costmap is not initialized

A valid layered costmap reference is needed to get the inscribed radius

Signed-off-by: Shrijit Singh <shrijitsingh99@gmail.com>

Co-authored-by: Shrijit Singh <shrijitsingh99@gmail.com>

* attempt at bringing up CI

* try adding SDF element for IMU initial_orientation_as_reference

* adding more initial orientation as reference

* turning initial_orientation_as_reference false

* adding tests for wp follower

* adding C++ 14 min requirement

* trying linking

* Adding minimum range parameter for raytrace clearing

* Comments in tests

* Uncrustify checks

* Modified default value of raytrace_min_range param and moved computation below plugin layer

* Tests for voxel layer raytracing

* Uncrustify checks

* Addressed pull request #2126 comments

* Using min range parameter for marking also

* Correcting build warnings

* Correct build fails

* Merging changes

* Adding comments

* Added unit tests for obstacle layer and addressed PR comments

* Addressed flake errors

* Flake8 errors

* Correcting W291 Flake8 error

* Variable name and comments changes

* Added raytracing params to the params files

* Modifying parameter names for obstacle marking

* Changed FollowWaypoints to follow_waypoints

* Updated the msg name as follow_waypoints

* Reversed changes to cmakelist

* Updated the use of follow_waypoints just as action name

* Replace follow_waypoints

* Update

* Updated ros info with the right syntax

* Debugging

* Add nav2_behavior_tree::BtActionServer

Signed-off-by: Sarthak Mittal <sarthakmittal2608@gmail.com>

* Fix cpplint errors

Signed-off-by: Sarthak Mittal <sarthakmittal2608@gmail.com>

* Remove unnecessary statements in BtActionServer

Signed-off-by: Sarthak Mittal <sarthakmittal2608@gmail.com>

* Make nav2_behavior_tree::BtActionServer a composable object

* Add comments

* Add on preempt callback, fix naming issues, and move tf to bt navigator

* Add separate implementation header for BtActionServer

* Fix cpplint error

* Pass plugin library names as argument to BtActionServer

* Remove action server getter and update onPreempt to not load BT

* Fix unnecessary include

* Fix function names

Signed-off-by: Sarthak Mittal <sarthakmittal2608@gmail.com>

* Fix typo

* Add particle cloud display rviz plugin

Signed-off-by: Sarthak Mittal <sarthakmittal2608@gmail.com>

* General refactoring

* Fix header guards

* Update bringup rviz files and default values

Co-authored-by: JF Dalbosco <jean-francois.dalbosco@safrangroup.com>

* Add speed limit filter

* Fix review items

* Remain only percent speed limit support

* Small review fixes

* Add error message for unknown cells on mask

* Fix missed items

* Ensure SpeedFilter is using accurate in time transform
instead of latest in TF Buffer

* Add costmap_filter_info_server into param description

* Set transform_tolerance to 0.1

* Update comment to be more informative

* Add DWB base obstacle critic test.

* Remove magic numbers.

* Fix build error.

* Fix out of bound for setting cost in costmap.

* Enable commented out code.

* include ZMQ publisher for Groot

very plain integration, should be made optionally through a launch parameter

* fix Groot crashing finding custom nodes in monitor mode

straight forward working fix. The manifest was missing, so Groot searched custom node IDs that it did not have. This is implemented correctly directly in BT.CPP V3 and should be used instead of an implementation in nav2_bt_engine

* refactor buildTreeFromText to createTreeFromText as in BT.CPP v3

* forward XML to createTreeFromText from BT.CPP v3 factory function

* Add createTreeFromFile forware to BT-factory function

* fix createTreeFromFile args..

* add personal copyright

I think this is okay for finding a nasty bug.. :)

* move creating ZMQ Publisher from run to dedicated function

this way the ZMQ Publisher ca be added to individual trees within the same factory. Should be important for switching trees (XML files)

* Add parameter for Groot Monitoring - default true. Also cleanup ZMQ

* Move haltAllActions() Implementation from .hpp to .cpp

* update Copyright in hpp of BT-engine

* make linters happy.. :)

* Update Groot parameter naming and chg default=0

* rename resetZMQGrootMonitor -> resetGrootMonitor

* add parameter to nav2_params.yaml - default = false

* add ZMQ params and logic for server/pub ports

* Fix RewrittenYaml ignoring Integers

Integers where converted as floats before which crashes get_parameter.. fun thing....

* add launch based tests for params and ZMQ

* Activate Dijkstra and A* switching tests, thanks to RewrittenYaml

* add pyzmq==19.0.2 via pip3 to CI test_workspace

* make flake8 linter happy

* make cpp linters happy

* add personal copyright

* add GoalUpdated BT node description in order to view the full default BT

only affects editor mode of Groot and not live monitoring

* make linter happy (unused import)

* remove unused groot-port replacement functions in test_system_launch.py

* add groot parameters to params.md

* get reloading BTs to work nicely with Groot

* pretty space for smac :)

* switch from unsinged to uint16_t

* fix converting string into float or int

* Revert "add pyzmq==19.0.2 via pip3 to CI test_workspace"

This reverts commit 7bca08121c88db3763771911e3c6b4c6f4f8ddeb.

* Switch to 4 spaces indent and other linter stuff for RewrittenYaml

* removed prints in test_system_launch.py

* linter stuff

* add python-zmq as test_depend in package.xml (instead of .CI_conf)

* enable groot monitoring by default

* remove ZMQ from naming (function / variable)

* remove variable zmq ports from testing scripts

* remove default ports in BT_engine, as they are set through (def-)params

* Remove complete test for "dynamic" ZMQ ports testing

* fix python-zmq depend location

* fix style

* swap missing Groot to default True

* fix rosdep zmq + flake8 fixes in system_tests

* remove debug logs + c_str()

* remove final debug_log

* Standardize time units per REP-103

* bond_timeout type

* Removing eloquent from build status table since its EOL

* Update README.md

* Update README.md

* adding unit testing for alignment util dwb critics

* adding twirling tests

* renaming test

* Setup for obstacle footprint test.

* Make std::string arguments const

* Added additional tests for obstacle footprint test.

* Add todo.

* Several code cleanup actions. Renamed the function p to getPoint. Removed commented out code. Fixed the trow->throw typo and explained some tests in comments.

* Test setting the parameters.

* Added the score tests.

* Use variables for parameters and add test for non-default values.

* Cleanup. Removed commented out code.