SBEL PhD student Huzaifa Mustafa Unjhawala, together with collaborators in the lab, has developed a new simulation tool, Chrono::CRM, designed to model deformable soils and vehicle–terrain interactions with high accuracy and efficiency. Chrono::CRM leverages Smoothed Particle Hydrodynamics (SPH) to represent terrain as a continuum, combined with optimized GPU implementations that enable simulations at unprecedented scales and speeds.
The simulator was validated against four physical experiments and one DEM-based numerical test, confirming its accuracy in predicting rover and vehicle behavior across diverse conditions. Demonstrations include NASA’s RASSOR robot performing autonomous digging and deposition, as well as real-time soil leveling controlled by adaptive blade positioning.
Chrono::CRM is fully integrated into Project Chrono, SBEL’s open-source multiphysics simulation framework. The tool enables large-scale simulations of up to 29 km terrain length on an NVIDIA H100 GPU, while still maintaining computational performance comparable to traditional semi-empirical models such as the Soil Contact Model (SCM).