DSAM: Dual-Strategy Simulation Acceleration-Based Model for Unmanned Ground Vehicle Mobility Prediction

The Next-Generation NATO Reference Mobility Model (NG-NRMM) currently integrates numerical simulation technologies based on the discrete element (DE) method and finite element (FE) method to predict the global mobility map of unmanned ground vehicles over soil terrains. However, computational efficiency remains a significant challenge. To address this issue, we propose an accelerated global mobility prediction method, DSAM. Unlike previous studies that focused solely on terrain modeling or computational acceleration, DSAM introduces a dual-strategy synergy that combines data-driven FE-DE coupling with adaptive soil patch computation to achieve order-of-magnitude simulation speedup. Specifically, our method enhances simulation efficiency through two primary strategies: First, in terrain modeling, we introduce a data-driven FE-DE coupling approach to overcome the computational dimensionality challenges inherent in DE simulations. Second, in simulation computation, we propose a non-complete moving soil patch method combined with multi-CPU parallel processing to substantially improve the real-time performance of mobility simulations. Experiments using real terrain data show that DSAM significantly improves prediction efficiency while maintaining accuracy.