Synthetic Environments for Multi-Physics Simulation of Robotic Operations in Complex Unstructured Environments

Robotics has the potential to transform physical labor across a wide range of domains, yet its deployment has remained largely confined to controlled environments such as factories, warehouses, and logistics centers. Expanding robotic operations to complex unstructured environments, including construction sites, disaster zones, natural terrain, and extraterrestrial surfaces, require advanced autonomy, real-time perception, and adaptive decision-making. Physical testing in these environments, or constructing physical replicas for testing, is expensive, hazardous, and difficult to scale. Digital representations offer an alternative, but they often lack the fidelity needed to capture the complexity of real-world conditions. This gap underscores the need for digital twins that can synthetically mimic real-world physics like model terrain, soil interactions, environmental forces, and dynamic robotic behavior for effective training and validation. The Industrialized Construction Innovation (ICI) team at the National Laboratory of the Rockies and the Robotic Space Exploration (RoSE) Lab at the Colorado School of Mines collaborated to develop and test synthetic digital twins for studying multi-physics interactions between robots and unstructured environments. This presentation will provide a high-level overview of this ongoing research, including how simulation tools such as NVIDIA Isaac Sim, Newton, Genesis, and ROS have been used to model and test robotic behavior in synthetic representations of lunar environments. It will also discuss how similar digital twins informed the deployment of robotic systems in real-world pilot studies, including the construction of a scaled energy infrastructure prototype, and how these developments can be extended to support robotic operations in future lunar missions.