Multi-Computer ROS Simulation Environment

Smalley, Chris, Chaomin Luo, Mohammad Utayba, Mohan Krishnan, and Mark Paulik

Simulation is a critical component for the development and testing of robotics applications. In this project a coordinated multicomputer simulation environment is developed using the Robot Operating System (ROS) and its associated simulation platform Stage. ROS is a widely used meta-operating system which is based on peer to peer communication architecture. Different processes and sensors can run on individual nodes and these nodes are able to communicate data directly with each other by publishing and subscribing to topics. For an autonomous robot to function at a high level it may need to run several different processes at the same time including image processing, local navigation, global navigation, mapping, etc. Some of the applications can be computationally intensive and a single computer is not likely to be able to run them all simultaneously along with a functioning simulation environment. For this reason a simulation environment capable of utilizing multiple computers to run applications in the same simulation environment was developed. Coordination of multiple computers running multiple applications within a simulation environment allows for a smoother transition to a similar system on the physical vehicle.