Noetic Compliance

[vinz-uts]

Hi, I'm trying to build the package on ROS Noetic (installed on Ubuntu 20.04), but a lot of errors are raised in the catkin_make process. Below an example of some errors' print:

/home/user/ros_ws/src/asv_wave_sim/asv_wave_sim_gazebo_plugins/src/Utilities.cc:154:44: error: ‘_sdf’ was not declared in this scope; did you mean ‘sdf’?
  154 | bool Utilities::SdfParamBool(sdf::Element& _sdf,
      |                                            ^~~~
      |                                            sdf
/home/user/ros_ws/src/asv_wave_sim/asv_wave_sim_gazebo_plugins/src/Utilities.cc:155:3: error: expected primary-expression before ‘const’
  155 |   const std::string& _paramName, const bool _defaultVal)
      |   ^~~~~
/home/user/ros_ws/src/asv_wave_sim/asv_wave_sim_gazebo_plugins/src/Utilities.cc:155:34: error: expected primary-expression before ‘const’
  155 |   const std::string& _paramName, const bool _defaultVal)
      |                                  ^~~~~
/home/user/ros_ws/src/asv_wave_sim/asv_wave_sim_gazebo_plugins/src/Utilities.cc:155:56: error: expression list treated as compound expression in initializer [-fpermissive]
  155 |   const std::string& _paramName, const bool _defaultVal)
      |                                                        ^
/home/user/ros_ws/src/asv_wave_sim/asv_wave_sim_gazebo_plugins/src/Utilities.cc:160:8: error: ‘size_t asv::Utilities::SdfParamSizeT’ is not a static data member of ‘class asv::Utilities’
  160 | size_t Utilities::SdfParamSizeT(sdf::Element& _sdf,
      |        ^~~~~~~~~
/home/user/ros_ws/src/asv_wave_sim/asv_wave_sim_gazebo_plugins/src/Utilities.cc:166:48: error: ‘_sdf’ was not declared in this scope; did you mean ‘sdf’?
  166 | double Utilities::SdfParamDouble(sdf::Element& _sdf,
      |                                                ^~~~
      |                                                sdf

I think the problem is related to the building process of the plugins files, maybe some change in the gcc compiler. Had anyone tryed to run the asv_wave_sim package on Noetic? Thanks for the support.

[nicolaje]

Hi @vinz-uts ,

I don't have any issue building the plugin on the development branch on Ubuntu20.04 with ROS Noetic.

Which branch are you building ?

[vinz-uts]

Hi, thanks for your reply. I got this error when I've tried to compile the master branch. Next I've downloaded the other branch and with the development branch I haven't also issues when I build the package. But in the last week I've always the same problem: the vehicle sinking down because there is some problem with the libHydrodynamicsPlugin. I got a lot of this print

[Err] [Wavefield.cc:896] point:  -7.63013 40.9786 0.359706
[Err] [Wavefield.cc:898] Water patch is too small
[Err] [Wavefield.cc:896] point:  -41.8836 -6.4903 0.359706
[Err] [Wavefield.cc:898] Water patch is too small

and, at the end:

[Err] [Model.cc:1155] Exception occured in the Init function of plugin with name[hydrodynamics] and filename[libHydrodynamicsPlugin.so]. This plugin will not run

I think I've do somthing wrong, but I don't understand what. I've downloaded the asv_wave_sim-develop package, I copy it in my ROS workspace and I've compiled it. In my package I've built my rov and I include in the <gazebo> tag the following code:

<!-- Hydrodynamics plugin -->
        <plugin name="hydrodynamics" filename="libHydrodynamicsPlugin.so">
            <!-- Wave Model -->
            <wave_model>ocean_waves</wave_model> <!-- before: ocean -->
            <!-- Hydrodynamics -->
            <damping_on>true</damping_on>
            <viscous_drag_on>true</viscous_drag_on>
            <pressure_drag_on>true</pressure_drag_on>
            <!-- Markers -->
            <markers>
                <update_rate>30</update_rate>
                <water_patch>false</water_patch>
                <waterline>false</waterline>
                <underwater_surface>false</underwater_surface>
            </markers>
        </plugin>

The package.xml file of my package is configured with

  <buildtool_depend>catkin</buildtool_depend>
  <build_depend>rospy</build_depend>
  <build_depend>std_msgs</build_depend>
  <build_depend>msg_generation</build_depend>
  <build_depend>actionlib_msgs</build_depend>
  <build_depend>uuv_gazebo_ros_plugins_msgs</build_depend>
  <build_export_depend>rospy</build_export_depend>
  <build_export_depend>std_msgs</build_export_depend>
  <exec_depend>rospy</exec_depend>
  <exec_depend>std_msgs</exec_depend>
  <exec_depend>message_runtime</exec_depend>
  <exec_depend>actionlib_msgs</exec_depend>
  <!--exec_depend>uuv_gazebo_ros_plugins_msgs</exec_depend-->
  <!--exec_depend>uuv_gazebo_ros_plugins</exec_depend-->
  <exec_depend>asv_wave_sim_gazebo_plugins</exec_depend>
  <exec_depend>gazebo_ros</exec_depend>
  <!--exec_depend>message_runtime</exec_depend-->

  <!-- The export tag contains other, unspecified, tags -->
  <export>
    <gazebo_ros gazebo_media_path="${prefix}" />
    <gazebo_ros gazebo_model_path="${prefix}/world_models:${prefix}/models" />
  </export>

I'm sure I don't understand how to correctly export the resources variables... Have you any idea to help me? Thanks in advice.

[srmainwaring]

@vinz-uts apologies for not getting back to you sooner. I can replicate the compile errors you are seeing on the master branch with ROS Noetic. These are caused by API changes between Gazebo9 and Gazebo11 and its dependencies on SDFormat.

The feature/fft_wave branch does compile and will run on ROS Noetic (and also ROS Melodic with custom installation of Gazebo11). This branch does not have the wave parameters exposed in the API - so is perhaps less useful for testing a variety of conditions.

The errors you are seeing are because the name of the ocean model in the feature/fft_wave branch has been shortened to ocean. If you change the element to read <wave_model>ocean</wave_model> in your model then you should see the buoyancy forces operating.

I have some maintenance to provide working branches for gazebo9, gazebo11, ignition-gazebo etc. I'll start with getting a version of master working with gazebo11.

[vinz-uts]

Thanks for your reply. I've tried to install the feature/fft_wave branch, then I've launched the ocean_world.launch file in the asv_wave_sim_gazebo package and the ocean model was correctly loaded. Then I've tried to copy this launch file in my package to verify that the path model is correctly exported and all works. When I try to launch my launch file, spawning my robot model, with this plugin in the <gazebo> tag

<plugin name="hydrodynamics" filename="libHydrodynamicsPlugin.so">
            <!-- Wave Model -->
            <wave_model>ocean</wave_model>
            <!-- Hydrodynamics -->
            <damping_on>true</damping_on>
            <viscous_drag_on>true</viscous_drag_on>
            <pressure_drag_on>true</pressure_drag_on>
            <!-- Markers -->
            <markers>
                <update_rate>30</update_rate>
                <water_patch>false</water_patch>
                <waterline>false</waterline>
                <underwater_surface>false</underwater_surface>
            </markers>
        </plugin>

a lot of this kind of errors are displayed on the terminal

[Err] [HydrodynamicsPlugin.cc:274] Wavefield Entity is null: ocean::wavefield_entity
[Err] [HydrodynamicsPlugin.cc:634] Wavefield is NULL

Have you any idea? I'm sure I'm doing some error. Thanks for your support!

[srmainwaring]

The SDF looks correct, it may be that you have another model called ocean in the Gazebo resource path that is being found first. This could come from the vrx or uuv_simulator packages if they are installed. If you expand the entity tree in the Gazebo side panel is the ocean model the one you are expecting (i.e. does it have the wave plugins attached)?

The other thing you could try is editing the ocean model and reverting it's name back to ocean_waves or similar, and update your boat models to match. That might help avoid conflicts with other packages.

[vinz-uts]

You were right, there were two model with the same name. I've tried to delete the ocean model folder from uuv_simulator instead of renaming the ocean folder in asv_wave_sim_gazebo folder. Now I'm sure that the ocean models loaded is correct (I've verified the info in the side panel of Gazebo) but I get the following errors:

[Err] [WavefieldSampler.cc:195] point:  9.06644 40.1202 0.359706
[Err] [WavefieldSampler.cc:197] Water patch is too small
[Err] [WavefieldSampler.cc:195] point:  -7.25091 41.0352 0.359706
[Err] [WavefieldSampler.cc:197] Water patch is too small
[Err] [WavefieldSampler.cc:195] point:  -7.63013 40.9786 0.359706
[Err] [WavefieldSampler.cc:197] Water patch is too small
[Err] [WavefieldSampler.cc:195] point:  -41.8836 -6.4903 0.359706
[Err] [WavefieldSampler.cc:197] Water patch is too small
[Msg] Water patch size: 0
[Err] [Model.cc:1155] Exception occured in the Init function of plugin with name[hydrodynamics] and filename[libHydrodynamicsPlugin.so]. This plugin will not run

and the vehicle is still sinking. Thanks a lot for your help!

[srmainwaring]

Ok - this is happening because the patch used to project the vehicle onto the water is not large enough to cover all the vertices in the vehicle mesh. (i.e. the AABB of the model projected onto the xy-plane must be covered by the patch). You can view the patches by setting <water_patch>true</water_patch> and pausing the simulation on start (let it evolve a couple of steps then stop before the model sinks).

Is the feature/fft_wave branch working with the models included in the repo? If these are working it may be that there is something amiss with the collision mesh on your model.

[vinz-uts]

I've enabled the water patch view setting <water_patch>true</water_patch>. I see the following is it a correct view? Previously I've used this same vehicle model (urdf file, mesh file, launch file, etc) on Melodic and Ubuntu 18 with the master branch of asv_wave_sim package and all works fine.

@edit If I use a cylinder instead of my mesh in the <geometry> tag in urdf, the '[Err] [WavefieldSampler.cc:197] Water patch is too small' error disappeared. But the [Err] [Model.cc:1155] Exception occured in the Init function of plugin with name[hydrodynamics] and filename[libHydrodynamicsPlugin.so]. This plugin will not run error remain and the vehicle sinks. I don't understand why the same mesh file had not problem with Melodic version.,

Thanks for your advices!

[srmainwaring]

I'd expect to see something more like this:

Or in wireframe mode:

(this on the feature/fft_waves branch).

Are the models in asv_wave_sim_gazebo/models working correctly? I've just set up an Ubuntu 20.04 VM with ROS Noetic as detailed here: https://github.com/srmainwaring/asv_wave_sim/issues/13#issuecomment-909197628, and am using that to generate the samples.

[vinz-uts]

The grid looks like it is splitted under the vehicle. I've tried to substituting the mesh files of my vehicle (both visual and geometry) with the duck model in the asv_wave_sim_gazebo package but it is still sinking (is this the correct way to use your models?). I don't understand why the same mesh file had not problem with Melodic version.

The [Err] [Model.cc:1155] Exception occured in the Init function of plugin with name[hydrodynamics] and filename[libHydrodynamicsPlugin.so]. This plugin will not run error remain and the object sinks.

[srmainwaring]

If your Gazebo resource paths are set up correctly then you can start the ocean world with

$ roslaunch asv_wave_sim_gazebo ocean_world.launch verbose:=true

Then navigate to the Insert tab on the Gazebo sidebar, select the duck model from the appropriate collection, and drop it into the ocean.

The water patch for the model doesn't look right, so there is something odd going on. I can't determine what is going on from the message, if you could post the entire console log (text not screenshot) that may help. Alternatively if you can share the model I can take a look at whether it works on my Ubuntu environment.

[vinz-uts]

Hi, I've found that the main problem was in my vehicle model. My mesh model was created in mm and Gazebo considered it in meters even if I used the scale option in my urdf <geometry> tag. I did not get the error abaut the waterpatch dimension anymore but the it was still visualized in the wrong way and, in addition, I started to have the same problem (low performance) reported by @nicolaje (I tried to use catkin_make --cmake-args -DCMAKE_BUILD_TYPE=RelWithDebInfo but performance doesn't improve). In the end, I replaced my vehicle geometry with a simpler shape (just a cylinder) and now it works. My vehicle geometry is probably too complex and involves too many calculations. I'll try to make it simpler.

Next I've switched to develop branch because I want to modify waves entity and others ocean's parameters. I was doing some tests with the coefficients of friction and damping.

<!-- Linear and Angular Damping -->
    <cDampL1>1.0E-6</cDampL1>
    <cDampL2>1.0E-1</cDampL2>
    <cDampR1>1.0E-6</cDampR1>
    <cDampR2>1.0E-6</cDampR2>

I don't understand how this parameters affect the behavior of the vehicle (the linear and quadratic coefficients seem to have the same effects). If possible, can you share a formula to better understand this relationship?

[srmainwaring]

If possible, can you share a formula to better understand this relationship?

Hi @vinz-uts, the two references to Jacques Kerner's blog at the bottom of the README contain the motivation for these parameters. They are the linear and quadratic coefficients for linear / rotational velocity dependent damping.

I'll close the issue as the main concern about building with ROS Noetic seems to be addressed.