Model submission for Marmotte (HD2) sensor configurations 1,2 by CTU-CRAS-Norlab

[peci1]

This is our sensor configuration of SuperDroid HD2 tracked robot. It is the same base platform as already existing MARBLE_HD2 robot.

• The robotic platform is the wide version of the "HD2 Treaded ATR Tank Robot Platform" from SuperDroid Robots
• The sensor rack sitting on top of the HD2 includes:
  • A Robosense RS-LiDAR-16 (or the new equivalent Leishen C16, depending on how many we break)
  • An omnivision setup of 6 Basler Ace2 (a2A1920-51gcPRO) cameras with C125-0418-5M-P Basler Lens
  • LED spotlights pointing in the same direction as the cameras
  • Xsens MTi-100 IMU
  • Rear-facing Intel Realsense D435i
  • The model includes the default SubT gas sensor
  • Weight: Approx. 52kg
  • HD2 platform with 4 motors and 2 batteries: 41kg
  • Sensor rack: 11 kg
• Sensor config 2 carries 12 comms breadcrumbs in addition to the above.

[peci1]

Some screenshots maybe useful during acceptance testing:

[peci1]

I downgraded model.sdf to SDF version 1.6 as requested for all our models. This model used no 1.7 features, so just changing the <sdf version> attribute was enough. I checked that with ign sdf --force-version 8.9.1 -k.

[acschang]

The initial assessment of the the model is complete however there are some issues to resolve before this model can be merged and utilized in a competitive setting:

• Please adjust the minimum distance clipping range on the D435 and Basler cameras to align with the standard clip range in the SubT repository.
• Please adjust the Omnicamera sensor to align with the versions accepted on the Spot and MARV platforms.
• Please add intrinsic values for all Basler cameras.
• Please adjust LIDAR parameters following the in-line comments to match the standard for LIDAR sensors in the SubT repository.
• The texture and color of the sensor payload on the vehicle almost looks like stone. Is this intended?

Please address the issues noted above in this pull request. Once the above issues are fixed, your submission will move onto the next review stage!

[peci1]

This is an example of the rendering bug which forced us to increase clipping distance to 0.4. Notice the (dis)appearing tracks in cameras based on the world orientation of the robot.

marmotte_cam_problem.mp4

[peci1]

I've solved some of the mentioned issues.

• Please adjust the minimum distance clipping range on the D435 and Basler cameras to align with the standard clip range in the SubT repository.
• not changed, there were rendering problems with lower clipping distances
• Please adjust the Omnicamera sensor to align with the versions accepted on the Spot and MARV platforms.
• Please add intrinsic values for all Basler cameras.
• Please adjust LIDAR parameters following the in-line comments to match the standard for LIDAR sensors in the SubT repository.
• The texture and color of the sensor payload on the vehicle almost looks like stone. Is this intended?
• It is intended to look like scratched metal, but due to our lack of experience with 3D modeling, it looks like marble.

I also added an RViz config for easy visual inspection of the sensor data.

[peci1]

I also moved battery plugin to the last place, following #901 (not sure if it's needed here, but it shouldn't hurt).

[acschang]

LGTM. Thank you for the clarification and changes.