search

SteveMacenski / spatio_temporal_voxel_layer

A new voxel layer leveraging modern 3D graphics tools to modernize navigation environmental representations
http://wiki.ros.org/spatio_temporal_voxel_layer
GNU Lesser General Public License v2.1
640 stars 189 forks source link

FIX-179 making STVL compatible with nav2 main branch #180

Closed ghost closed 3 years ago

ghost commented 4 years ago

Build is successful though I am yet to test it.

ghost commented 4 years ago

This is for the issue: https://github.com/SteveMacenski/spatio_temporal_voxel_layer/issues/179

ghost commented 4 years ago

Thank you Steve. I'll raise another PR tomorrow.

SteveMacenski commented 4 years ago

You can just add your commit to this one, you don't need to have another PR

ghost commented 4 years ago

Hey @SteveMacenski Can you please review the changes again? I'm still observing the same behaviour I observed in https://github.com/SteveMacenski/spatio_temporal_voxel_layer/issues/181 .

I think this is happening because I am not using the correct clock source. Can you help me out on this?

ghost commented 4 years ago

I'm probably missing something very trivial.

I'm getting the following messages on my terminal:

[controller_server-4] [WARN] [1600514543.907029725] [local_costmap.local_costmap]: /camera/depth/color/points buffer updated in 17.51s, it should be updated every 0.00s.
[controller_server-4] [WARN] [1600514543.907114075] [local_costmap.local_costmap]: /camera/depth/color/points buffer updated in 17.51s, it should be updated every 0.00s.
[planner_server-5] [WARN] [1600514543.958423666] [global_costmap.global_costmap]: /camera/depth/color/points buffer updated in 16.42s, it should be updated every 0.00s.
[planner_server-5] [WARN] [1600514543.958509501] [global_costmap.global_costmap]: /camera/depth/color/points buffer updated in 16.42s, it should be updated every 0.00s.
[controller_server-4] [WARN] [1600514544.007036231] [local_costmap.local_costmap]: /camera/depth/color/points buffer updated in 17.61s, it should be updated every 0.00s.
[controller_server-4] [WARN] [1600514544.007119344] [local_costmap.local_costmap]: /camera/depth/color/points buffer updated in 17.61s, it should be updated every 0.00s.

when I set expected_update_rate to 1.0

ghost commented 4 years ago

On further investigation and testing I realized that the changes in this PR are fine and working.

The issue was with the realsense ROS2 ~driver~ wrapper that I was using. https://github.com/intel/ros2_intel_realsense driver does not work with this plugin (although it does publish the pointcloud) while https://github.com/IntelRealSense/realsense-ros works just fine.

I tested on ros2 foxy navigation main branch with ubuntu 20.04

SteveMacenski commented 4 years ago

What do you mean its the driver? I'm not sure how that could be possible. But if it works with another driver (or a 3rd one), then yeah, it might be that.

ghost commented 4 years ago

What do you mean its the driver? I'm not sure how that could be possible. But if it works with another driver (or a 3rd one), then yeah, it might be that.

Sorry, my bad. I meant the ROS2 wrapper not the driver.

SteveMacenski commented 4 years ago

In what specific way does it not work? What about the PointCloud2 message or TF2 not work?

ghost commented 4 years ago

In what specific way does it not work? What about the PointCloud2 message or TF2 not work?

https://github.com/SteveMacenski/spatio_temporal_voxel_layer/issues/181 This is the behaviour I observed.

SteveMacenski commented 4 years ago

But why was that happening as a result of these ROS2 wrappers?

ghost commented 4 years ago

But why was that happening as a result of these ROS2 wrappers?

I'm not sure why. Haven't investigated that yet. Will inform you about whatever I discover.

SteveMacenski commented 4 years ago

OK, if you look at a message from both, it should be pretty clear. If you use the --no-array you can just see the headers / content without all the console spam. Maybe a timing issue like you mention, maybe a frame transformation issue.

I'd like us to know exactly why this has happened before merging any patch.

ghost commented 4 years ago

OK, if you look at a message from both, it should be pretty clear. If you use the --no-array you can just see the headers / content without all the console spam. Maybe a timing issue like you mention, maybe a frame transformation issue.

Sure. I will do that once I am in office (where the camera is).

I'd like us to know exactly why this has happened before merging any patch.

Sure.

SteveMacenski commented 4 years ago

@garyrightbot any update? There is just a couple really small things before we can merge this in.

ghost commented 4 years ago

@SteveMacenski I will do this within a week. Currently, I do not have access to the camera.

SteveMacenski commented 3 years ago

Superseded by #191