Update dwb_controller as per robot_navigation stack.

[shivaang12]

---

Basic Info

Info	Please fill out this column
Ticket(s) this addresses	#229
Primary OS tested on	Ubuntu 18.04
Robotic platform tested on	Turtlebot3

---

Description of contribution in a few bullet points

• Ported most of the changes from the robot_navigation repo.

---

Future work that may be required in bullet points

• I haven't ported some changes which includes input_params and some minor changes.
• I am planning to complete this work shortly.
• Need to know if anyone wants me to port input_params which I think is redundant. But it is just my thoughts.

[SteveMacenski]

Can you describe what input params are?

Thanks for this, great stuff. Have you tested this in simulation with the turtlebot? Also, I see you added a few parameters, can you make sure to update those in the nav2_params.yaml and in the website docs.

This is a big help, thanks for your hard work

[shivaang12]

I have tested this on my old repo (which is approx 3 weeks old) as I am having issues with running it on current nav stack (I still working on the issue). It would be helpful if anyone volunteers to check on their system till I fix my system.

Below are the input params
nav_grid - A templatized interface for overlaying a two dimensional grid on the world. (Which has not been incorported in nav2 stack)
start_pose - Start pose.
goal_pose - Goal pose .
velocity - velocity.
which are being published in the new update.

I will update the params nav2_params.yaml.

[SteveMacenski]

Looks like a good first run at it.

How has the testing for this been going? Should we update the default parameters?

[SteveMacenski]

Pull in master, CI is fixed

[shivaang12]

@SteveMacenski Would it okay if I submit new PR for the updates to doc for new params. As my fork do not have "configuration" udpate in "sphinx_doc".

[SteveMacenski]

please add them in here- its one of those technical debt things that we shouldn’t get in the habit of ignoring. Plus when this goes in, it’ll make the documentation on the website outdated and we should keep that current.

Also, you'll need to pull in current changes in order for CI to run and succeed.

[shivaang12]

Okay! I will have to rebase then.

[SteveMacenski]

Make sure to touch these files too:

https://github.com/ros-planning/navigation2/tree/f2170be344f4b7f875632fcf5210381cff3f41c4/sphinx_doc/configuration/dwb-plugins

https://github.com/ros-planning/navigation2/blob/f2170be344f4b7f875632fcf5210381cff3f41c4/sphinx_doc/configuration/configuring-the-controller.rst

[SteveMacenski]

Toggled to retrigger CI

[codecov]

Codecov Report

Merging #1501 into master will decrease coverage by 0.12%.
The diff coverage is 27.1%.

@@            Coverage Diff             @@
##           master    #1501      +/-   ##
==========================================
- Coverage   37.47%   37.34%   -0.13%     
==========================================
  Files         226      227       +1     
  Lines       11774    11881     +107     
  Branches     5084     5149      +65     
==========================================
+ Hits         4412     4437      +25     
- Misses       3985     4034      +49     
- Partials     3377     3410      +33

Flag	Coverage Δ
#project	37.34% <27.1%> (-0.13%)	⬇️

Impacted Files	Coverage Δ
...ns/include/dwb_plugins/limited_accel_generator.hpp	100% <ø> (ø)	⬆️
...controller/dwb_core/include/dwb_core/publisher.hpp	50% <ø> (ø)	⬆️
...dwb_critics/include/dwb_critics/rotate_to_goal.hpp	100% <ø> (ø)	⬆️
...ns/include/dwb_plugins/standard_traj_generator.hpp	100% <ø> (ø)	⬆️
...lugins/include/dwb_plugins/simple_goal_checker.hpp	100% <ø> (ø)	⬆️
...er/dwb_core/include/dwb_core/dwb_local_planner.hpp	100% <ø> (ø)	⬆️
...2_dwb_controller/dwb_plugins/test/goal_checker.cpp	50% <0%> (-0.95%)	⬇️
nav2_core/include/nav2_core/goal_checker.hpp	66.66% <0%> (-33.34%)	⬇️
...2_dwb_controller/dwb_core/src/trajectory_utils.cpp	0% <0%> (ø)
...roller/dwb_plugins/src/limited_accel_generator.cpp	56.25% <100%> (+8.25%)	⬆️
... and 16 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update aaa9790...7f0572e. Read the comment docs.

[SteveMacenski]

Looks like you have a bunch of linting problems

[SteveMacenski]

LGTM after linting issues and passes tests

Thanks for your attention to detail

[SteveMacenski]

Do you notice if this is still true after your updates? #938

[shivaang12]

Haven't noticed this behavior but let me check.

[shivaang12]

Nope! This update does not solve the issue. Robot is still traversing dangerously close to obstacles.

[SteveMacenski]

Got it.

With these updates should any of the gains be tuned differently? (Ei is control worse or changed from before the update?)

[shivaang12]

This update only address some calculation (algorithmic) bugs/improvements not much on gains.

[SteveMacenski]

I know. If you change algorithms, that is going to inherently change how the current gains are going to effect the system.

Can you do some testing to show that this either improves controller behavior or at least just doesn't effect it? If it's worse, updating the gains to be at minimum as good as today?

[SteveMacenski]

I did some testing, I'm happy with the performance

Thanks for your hard work in getting these updates in!

[shivaang12]

Cool! Thank you for testing (on my behalf).