Enhanced Multiscale Off-Road Mobility Prediction Capability with Machine Learning Constitutive Modeling for Large Deformable Granular Terrain

Project began in 2020 and was complete Q1 2023.

A high-fidelity off-road mobility model is an integral part of the Next-Generation NATO Reference Mobility Model (NG-NRMM) to enable reliable operational planning. Whereas discrete-element (DE) terrain models have been utilized to predict granular soil behavior, the computational cost increases prohibitively high as the number of DE particles increases to consider a wider range of terrain; especially when constructing stochastic mobility maps, requiring a large number of simulation runs at different locations in the map for different soil at different speeds to predict distribution of “speed-made-good”.

The primary objective of this study is to develop a new hierarchical multiscale off-road mobility modeling procedure, enhanced by Material Point Method (MPM) and the machine learning (ML) soil constitutive modeling for the lower-scale representative volume element (RVE), to eliminate limitations of the existing finite element discrete element (FE-DE) multiscale and single-scale DE terrain models. The proposed model will enable quick and reliable vehicle mobility prediction on large deformable granular terrain in challenging missions.

This work builds upon the PI and his team’s previous ARC project on the FE-DE multiscale off-road mobility model. The new contributions of this work lie in the following points:

• New development of the ML soil constitutive model and its training strategies for lower-scale RVEs to speed up online RVE calculations for large deformable granular terrain
• New development of the MPM macro-scale terrain model for off-road mobility simulation to eliminate the limitation of the FE macro-scale model in predicting large terrain deformation
• New development of the MPM-ML multiscale off-road mobility simulation framework with scalable parallel computing scheme to enable quick and reliable mobility prediction on large deformable granular terrain

Journal Papers:

1. Chen, G., Yamashita, H., Ruan, Y., Jayakumar, P., Gorsich, D., Knap, J., Leiter, K. W., Yang, X., and Sugiyama, H., 2023, “Hierarchical MPM-ANN Multiscale Terrain Model for High-Fidelity Off-Road Mobility Simulations: A Coupled MBD-FE-MPM-ANN Approach”, ASME Journal of Computational and Nonlinear Dynamics, vol. 18, pp. 071001-1-13. DOI:10.1115/1.4062204
2. Chen,G.,Yamashita,H.,Ruan,Y.,Jayakumar,P.,Knap,J.,Leiter,K.W.,Yang,X.,andSugiyama,H.,2021, “Enhancing Hierarchical Multiscale Off-Road Mobility Model by Neural Network Surrogate Model”, ASME Journal of Computational and Nonlinear Dynamics, vol. 16, pp. 081005-1-12. DOI:10.1115/1.4051271
3. Yamashita, H., Chen, G., Ruan, Y., Jayakumar, P. and Sugiyama, H., 2020, “Parallelized Multiscale Off- Road Vehicle Mobility Simulation Algorithm and Full-Scale Vehicle Validation”, ASME Journal of Computational and Nonlinear Dynamics, vol. 15, pp. 091007-1-14. DOI:10.1115/1.4046666

Conference Papers:

1. Chen, G., Yamashita, H., Sugiyama, H., Ruan, Y., Jayakumar, P., Gorsich, D., Leiter, K. W., Knap, J. and Yang, X., 2022, “Modeling Large Deformable Terrain with Material Point Method for Off-Road Mobility Simulation”, Proceedings of ASME International Conference on Multibody Systems, Nonlinear Dynamics, and Control (ASME DETC2022-89632), St. Louis, MO, United States. DOI:10.1115/DETC2022-89632
2. Chen,G.,Yamashita,H.,Ruan,Y.,Jayakumar,P.andSugiyama,H.,2020,“MultiscaleOff-RoadMobility Simulation with Computational Load Balancing for Lower-Scale Discrete-Element Models”, Proceedings of ASME International Conference on Multibody Systems, Nonlinear Dynamics, and Control (ASME DETC2020-22195), online. DOI:10.1115/DETC2020-22195

#3.17