A Physics-Based Continuum Model for Versatile, Scalable, and Fast Terramechanics Simulation

2025-07-07

AI Summary

To address the low fidelity of empirical models and the prohibitive computational cost of discrete element method (DEM) simulations in complex terramechanics, this paper proposes Chrono::CRM—a physics-driven continuum model based on smoothed particle hydrodynamics (SPH). Integrated into the open-source Chrono multibody dynamics engine, Chrono::CRM introduces GPU acceleration and an innovative “active domain” dynamic particle management strategy. It enables high-fidelity simulation of rigid–soft coupling and deformable wheel/tracked-vehicle–terrain interactions. Experimental validation on realistic scenarios—including NASA’s MGRU3 lunar rover—demonstrates that Chrono::CRM achieves excavation and bulldozing accuracy comparable to DEM while delivering computational efficiency on par with semi-empirical models. The framework scales to terrain domains up to 10 km in extent and supports simulations with over 100 million particles, enabling near-real-time large-scale terrain dynamics simulation.

Technology Category

• Planning, Routing, and Scheduling: Mixed Discrete/Continuous Planning
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior
• Intelligent Robots: Motion and Path Planning

Application Category

• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Vertical and domain-specific search

Abstract

This paper discusses Chrono's Continuous Representation Model (called herein Chrono::CRM), a general-purpose, scalable, and efficient simulation solution for terramechanics problems. Built on Chrono's Smoothed Particle Hydrodynamics (SPH) framework, Chrono::CRM moves beyond semi-empirical terramechanics approaches, e.g., Bekker-Wong/Janosi-Hanamoto, to provide a physics-based model able to address complex tasks such as digging, grading, as well as interaction with deformable wheels and complex grouser/lug patterns. The terramechanics model is versatile in that it allows the terrain to interact with both rigid and flexible implements simulated via the Chrono dynamics engine. We validate Chrono::CRM against experimental data from three physical tests, including one involving NASA's MGRU3 rover. In addition, the simulator is benchmarked against a high-fidelity Discrete Element Method (DEM) simulation of a digging scenario involving the Regolith Advanced Surface Systems Operations Robot (RASSOR). Being GPU-accelerated, Chrono::CRM achieves computational efficiency comparable to that of semi-empirical simulation approaches for terramechanics problems. Through an ``active domains'' implementation, Chrono::CRM can handle terrain stretches up to 10 km long with 100 million SPH particles at near interactive rates, making high-fidelity off-road simulations at large scales feasible. As a component of the Chrono package, the CRM model is open source and released under a BSD-3 license. All models and simulations used in this contribution are available in a public GitHub repository for reproducibility studies and further research.

Problem

• Develops physics-based terramechanics model for complex tasks
• Validates model with experimental data and DEM benchmarks
• Achieves scalable, efficient simulation with GPU acceleration

Innovation

• Physics-based model for versatile terramechanics simulation
• GPU-accelerated with active domains for scalability
• Open-source BSD-3 licensed for reproducibility