carla-autoware
carla-autoware copied to clipboard
"Failed transform from base_link to ego_vehicle/lidar/lidar1" and autoware does not seem to connect properly
Hi,
I followed the setup and run the agent instructions of the branch joel-mb/package_0_9_10
and modified run.sh
as discussed in #92 .
Everything seems fine until when the rviz window pops up. A screenshot is shown below:
The only error message that ever appeared in the docker window is:
[ERROR] [1599623780.456602301, 52.439418315]: Failed transform from base_link to ego_vehicle/lidar/lidar1
Could you give me some hints on what goes wrong here?
Also experiencing this problem, hoping to find the answer.
Hi @AIasd @AminoACID123. Can you try again following the new README instructions?
I am getting the same issue, I fallowed the README instructions @joel-mb
@joel-mb I used the updated Dockerfile to generate the docker image and everything works fine for me now. Thanks.
Hi I am having the same problems, and the ego_vehicle is not moving at all. Could you please post solution if you managed to get it working? thanks
@joel-mb I used the updated Dockerfile to generate the docker image and everything works fine for me now. Thanks.
Still the same :(
Based on my experiences, here is my advice:
- Use the newest Dockerfile.
- Check the integrity of the
autoware-contents
folder. - If it still doesn't work, check the log files located in
~/.ros/log/
folder for detailed information.
thanks, I just got it working. I followed a comment from Felix-Feng-90 (can't find that comment anymore in here), basically I forgot to install Git lfs (silly mistake). After installing it, I repeated the whole process to build carla-autoware docker, alongside carla 0.9.10.1 docker, and everything works well for me
@Jasonzhang11111 Glad to hear that installing Git lfs also works for you. I removed the comment because I found that although it works, there are still plenty of warning messages (especially some "topics not subscribed ones"), was trying to fix them as well. However, at the moment, I only solved the "state publisher" is deprecated warning, which needs to change the type of robot_state_publisher inside vehicle_model.launch
Still the same with git-lfs
thanks, I just got it working. I followed a comment from Felix-Feng-90 (can't find that comment anymore in here), basically I forgot to install Git lfs (silly mistake). After installing it, I repeated the whole process to build carla-autoware docker, alongside carla 0.9.10.1 docker, and everything works well for me
Following the instruction I can view point clouds and the image in YOLO box. However, I still had the error Failed transform from base_link to ego_vehicle/lidar/lidar1. And the ego_vehicle did not move anymore even though I set the 2D Nav goal. Does someone encounter the problem and solve it?
Hi @Basavaraj-PN @7n42Oaq1 ,
Are you able to see the vector map and the point cloud map in Rviz? As others have said, this may be related to not having the maps correctly cloned. Please check that inside the autoware-contents
folder you have all the *.pcd
, *csv
and *.osm
files.
Apart from this, can you check if there is any node crashing when running the agent?
@joel-mb I have attached terminal output and autoware-contents,
Terminal log -> launch_log.txt
At line num 2740 (log file) there is one error , terminate called after throwing an instance of 'boost::exception_detail::clone_impl<boost::exception_detail::error_info_injector<boost::lock_error> >' what(): boost: mutex lock failed in pthread_mutex_lock: Invalid argument
Autoware_Contents
I verified all *.pcd, *.csv and *.osm files exists.
@joel-mb my Rviz is shown like this
I think everything is ok except the vehicle does not move.
@joel-mb my Rviz is shown like this
I think everything is ok except the vehicle does not move.
@joel-mb I found my problem that I modified the Dockerfile to compile ros packages in ros-bridge. I repeated the whole process to build carla-autoware docker follow the readme and now everything works. Thanks.
@7n42Oaq1 glad to here is working now for you!
@Basavaraj-PN I realized that the vision_darknet_detect
node is also dying. Can you check what happends if you disable the vision nodes? Commenting the vision_detection
include tag in my_detection.launch
should be enough.
Hello all,
I installed git-lfs and repeated the whole process again.
Be patient when doing git clone, it might hang a while.
Then it works perfectly!
I installed git-lfs and follow the read me to build the docker . but still get the error "Failed transform from base_link to ego_vehicle/lidar/lidar1", the ego vehicle still can not move
I didn't use Docker to build the env and met the same issue. All is right unless the vehicle didn't move. I got the warning below:
[ WARN] [1630580298.268371780, 99.252712106]: Waiting for current_pose topic ...
[ WARN] [1630580298.268480414, 99.252712106]: Waiting for current_velocity topic ...
I didn't use Docker to build the env and met the same issue. All is right unless the vehicle didn't move. I got the warning below:
[ WARN] [1630580298.268371780, 99.252712106]: Waiting for current_pose topic ... [ WARN] [1630580298.268480414, 99.252712106]: Waiting for current_velocity topic ...
pip install simple-pid pygame networkx==2.2
This fixed my issue.
I installed git-lfs and repeated the whole process again.it is success. thank you @KailinTong
Hello all,
I installed git-lfs and repeated the whole process again.
Be patient when doing git clone, it might hang a while.
Then it works perfectly!
Thank you @KailinTong !!!
Hi, can someone please tell me the step-by-step process to run Carla and autoware docker? I did the following (Ubuntu20.04):
- Installed Docker 19.03
- Installed Nvidia Docker
- Installed Carla Docker (0.9.10.1)
- Installed Git-lfs (and checked that it was working with
$ git lfs install
) - Cloned Carla Autoware Repository
$ git clone --recurse-submodules https://github.com/carla-simulator/carla-autoware
- Built Carla autoware image to generate a
carla-autoware:latest
docker image - Run on one terminal the Carla server
- Run the carla-autoware image
- Start the agent by using:
$ roslaunch carla_autoware_agent carla_autoware_agent.launch town:=Town03
- Rviz is showing the vehicle on one part of the map and the sensing suite in another. Camera is not working.
Error appears to be vehicle model, No transform from [base_link] to [world]
I did have to use sudo to run the autoware docker image: $ sudo ./run.sh
Hi, which cuda version you are using?