Synthesizing Executable Simulations from Structural Models of Component-Based Systems

Abstract

Experts in robotics systems have developed substantial software tools for simulation, execution, and hardware-in-the-loop testing. Unfortunately, many of these robotics-domain software infrastructures pose challenges for a robotics expert to use, unless that robotics expert is also familiar with middleware programming, and the integration of heterogeneous simulation tools. In this paper, we describe a novel modeling language designed to bridge these two domains in an intuitive visual representation. Using this metamodel-defined modeling language, we can design and build structural models of robotics systems, and synthesize experiments from these constructed models. The restrictions implicit (and explicit) in the visual language guide modelers to build only models that can be synthesized, a "correct by construction" approach. We discuss the impact of this language with a running example of an autonomous ground vehicle, and the hundreds of configuration parameters and several simulation tools that are necessary in order to simulate this complex example.