ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange |
1 | initial version |
For ROS 1, I wouldn't recommend running the master node off board the robot platform, as wifi network drops and node restarts can really throw a wrench into the system. You could install multimaster in a separate docker image, then push that image to the robot and use it to start a container there. You'll want to check how the other containers are attached to the network, my guess is via --net=host
driver, so that your multimaster can connect to both the other containers, as well as IP addresses external to the robot. That said, I haven't used multimaster much, as it always seemed rather fragile.
You could try combining ROS 2 with ROS 1 using the ros1_bridge to link robots together. Then you could also control the white list of topics and interfaces to bridge across robots. I think this is similar to what clearpath does internally for their multi robot deployments. However, if you rely on a lot of custom HSR message types, you'd have to build the ros1_bridge from source: