Following the successful release 1.0.0 of Autoware.Auto we have explored how we can pair it with a Kubernetes setup to easily distribute and manage the different modules.

To showcase the simplicity of the distributed approach we have taken the 3D Perception stack demo and split it into 3 Kubernetes deployments:

• One to replay the Velodyne pcap data.
• One responsible for the sensing module, which in this case kicks off the front and rear Velodyne drivers.
• The last one to run all the 3D Perception stack nodes, namely point cloud filter transform, ray ground classifier and the euclidean cluster.

To get it closer to what a working configuration on a vehicle would look like we have created individual Docker images for each deployment where the minimal set of nodes have been built, so they are lighter than a full Autoware.Auto image.

As for the hardware we are using AutoCore’s PCU to run the Autoware.Auto work and a laptop for visualization purposes.

To get the full step-by-step blog post is here.