Scalable Interactive Autonomous Navigation Simulations on HPC
Brewer, Wesley (DoD HPCMP PET/GDIT)
Intersection of Digital Engineering and High Performance Computing/High End Computing
We present our work of enabling scalable interactive autonomous navigation simulations on HPC. The entire workflow consists of many different software components deployed within Singularity containers and communicating using both the Robotic Operating System's (ROS) publish subscribe and the Message Passing Interface (MPI). We used the Singularity container networking interface (CNI) to enable virtual networking within the containers, so that multiple containers can run the various components using different IP addresses on the same compute node. The Virtual Autonomous Navigation Environment (VANE) communicates with the Autonomous Navigation Virtual Environment Laboratory (ANVEL) over ROS and is distributed across multiple compute nodes via MPI along with OpenMP for shared memory parallelism within each compute node. The user interfaces with the navigation environment using an XFCE desktop with virtual workspaces running over a noVNC-VNC containerized deployment through a double-hop ssh tunnel. We automate the complete launch process using a custom iLauncher plugin. We report our findings of benchmarked scalable performance with multiple vehicle simulations. Finally, we discuss our most recent progress in dynamic grass modeling in CUDA.