auto exposure limit [ RS2_OPTION_AUTO_EXPOSURE_LIMIT,] clarification request

[AndreV84]

Could you clarify how to get it work? The example below did not work, as per dev's feedback

rs2_set_option((rs2_options*) sensor, RS2_OPTION_AUTO_EXPOSURE_LIMIT, 8000, &e);

[AndreV84]

update : seems it may be firmware version related issue
reference url: #10241 (comment)
but devs said they tried the latest

[MartyG-RealSense]

Hi @AndreV84 Please see my responses at #10241 (comment) in regard to this question. Thanks!

[MartyG-RealSense]

Hi @AndreV84 Do you require further assistance with this case, please? Thanks!

[AndreV84]

@MartyG-RealSense
Thank you for following up
the suggestion was to try with downgraded firmware
or with the development branch or with both? or only one of the two?

[MartyG-RealSense]

Both suggestions are valid, though I would recommend trying the firmware downgrade first as it is the quickest and easiest of the two methods.

[MartyG-RealSense]

Hi @AndreV84 Do you require further assistance with this case, please? Thanks!

[AndreV84]

@MartyG-RealSense
We haven’t tested it yet
Once the remote team updates I shall get update you in this regard
Thanks

[MartyG-RealSense]

Thanks very much for the update!

[MartyG-RealSense]

Hi @AndreV84 Do you still require assistance with this case, please? Thanks!

[AndreV84]

@MartyG-RealSense
consider the ticket closed, please

[MartyG-RealSense]

Thanks very much @AndreV84 :)

[AndreV84]

@MartyG-RealSense
On one of jetsons I've build the sdk with rsusb=on
cmake -DFORCE_RSUSB_BACKEND=ON -DBUILD_PYTHON_BINDINGS:bool=true -DPYTHON_EXECUTABLE=/usr/bin/python3 -DCMAKE_BUILD_TYPE=release -DBUILD_EXAMPLES=true -DBUILD_GRAPHICAL_EXAMPLES=true -DBUILD_WITH_CUDA:bool=true -DBUILD_WITH_CUDA:bool=true -DCMAKE_CUDA_ARCHITECTURES=72 -DCMAKE_CUDA_COMPILER_ENV_VAR=/usr/local/cuda-11.4/bin/nvcc -DCMAKE_CUDA_COMPILER=/usr/local/cuda-11.4/bin/nvcc -D OpenGL_GL_PREFERENCE=GLVND ..

Connected devices:
 23/03 04:58:25,660 WARNING [70368799660288] (messenger-libusb.cpp:42) control_transfer returned error, index: 300, error: Resource temporarily unavailable, number: b
1) Name: Intel RealSense D455

comments about the warning?

[MartyG-RealSense]

In the moment that the warning occurs, there may be a communication problem between the camera and the Jetson. If the warning only occurs a small number of times during launch and then stops, or generates once or twice per minute after launch, then it can be safely ignored.

[AndreV84]

@MartyG-RealSense

https://github.com/NVIDIA-ISAAC-ROS/.github/blob/main/profile/realsense-setup.md
`I'm m trying to run nvidia container with realsense d457

it as inside the same sdk version
but it won't list the camera from inside of the container while it will list the camera from outside of the container

[MartyG-RealSense]

It may be worth first looking at the general non-RealSense video in the link below about using a GMSL camera in Docker with Nvidia.

https://www.youtube.com/watch?v=-swTFYO3Fuc

[AndreV84]

I do not see where to procure that specific autoware-docker [ as in the video] deployment to test with it
how from nvidia realsense docker we can query camera or camera module from terminal to determine if it is detected by the system or not? thanks

[AndreV84]

trying with autoware
do you suggest the realsense camera will work with autoware with or without docker?

[MartyG-RealSense]

There are no previously reported cases of RealSense being used with Autoware, so I do not have advice to offer about that.

Autoware.Auto can apparently be used with Docker though.

https://idorobotics.com/2021/08/31/step-by-step-autoware-auto-installation-guide/

[AndreV84]

I managed to setup autoware
https://autowarefoundation.github.io/autoware-documentation/main/tutorials/ad-hoc-simulation/rosbag-replay-simulation/
but I do not see any clue how to integrate gmsl into the setup

moreover I have found a separate concerned in gmsl segment of tier4 that supports jetson
https://github.com/tier4/perception_ecu_container
but haven't managed to deploy it
neither see a clue how to get gsml d457 outputs with it

[MartyG-RealSense]

I have highlighted your question about using D457 in Docker to my Intel RealSense colleagues.

[AndreV84]

[MartyG-RealSense]

Thanks very much @AndreV84 for the image! Were you able to access your D457 in an Autoware Docker container using the above information?

My colleagues have advised that D457 has not been tested in Docker and if it cannot be detected in Docker then do not use it in Docker. They also recommended approaching Nvidia if you require further advice on this subject.

[AndreV84]

the chat assumption is that there is a camera launch file + camera driver for ros that generates camera topic [ perhaps like v4l topic]
the nvidia limitation for running vslam is to use their container with humble docker
will have to investigate it further

[AndreV84]

what are the chances that installing udev rules with d455 connected could result in faulty occurance sometimes? does it make sense to reinstall once installed udev rules that were installed while d455 connected m but to reinstall it again with d455 disconnected?

[MartyG-RealSense]

I would think that having one particular camera model attached during udev rule installation would not negatively affect how the udev rules handle other models, as the same set of udev rules handle all RealSense camera models that are plugged in.

[AndreV84]

which ros node you could suggest to generate topics from d457? maybe you could share recorded rosbag with such topics for playing them for experimenting with autoware? Thanks

[AndreV84]

does ros realsense node without docker generate gmsl camera topics?
@MartyG-RealSense

[MartyG-RealSense]

There is currently no information about D457 support for GMSL mode. In USB mode it should behave like a D455 and be able to generate topics.

I will highlight your question about D457 GMSL support to my colleagues on the RealSense ROS development team.

[MartyG-RealSense]

The RealSense ROS team confirmed that D457 support has already been merged into the ros2-development wrapper in January 2023 by IntelRealSense/realsense-ros#2557 and confirmed to work with rs_launch.py

You should be able to check which topics are published in the wrapper by a D457 on a GMSL connection with ros2 topic list