Issue with SLAM and Fusion example

Hello, I am using the docker image for the SLAM node, we are attempting to evaluate the mutli-sensor system and I am modeling it off the example which uses the robot_localization package to use multiple instances of the SLAM program.

The problem is the robot_localization package isn’t included into the docker image.(as far as I can tell) and when I follow the steps here:

https://nvidia-isaac-ros.github.io/getting_started/isaac_ros_buildfarm_cdn.html

My docker image gets messed up, apparently it resets at shutdown as when I restart it it’s like it was never there(which is a good thing as it erased my ros2 install).

I understand I can -commit the docker image to be used in the future but like I mentioned when I followed the steps above it ended up destroying my ros2 install.

Any suggestions?

My local ROS2 installation can find and install the localization package no problem so I suppose I could spawn the node locally but I’m not sure if that would work or not.

I just tested and sudo apt install ros-humble-robot-localization works inside the container.

Make sure you have the latest version of isaac_ros_common.

If running that command works inside the container, you can create a new docker file and add it to the config file as specified here.

1 Like

Thank you @reza11

yes, you can install the binaries of robot-localization without building from the source just running from the docker container:

sudo apt install ros-humble-robot-localization

When I run this I get an error:

Reading package lists… Done
Building dependency tree
Reading state information… Done
E: Unable to locate package ros-humble-robot-localization

I may have built the docker image, I don’t remember exactly. I’m currently running ubuntu 22.04

Which device are you using?

Please check if you are working with Jetpack 5.1.2 and working with the Isaac ROS docker on Isaac ROS common.