A substantial body of terramechanics research has been performed at the U.S. Army Tank Automotive Research, Development, and Engineering Center (TARDEC) and the U.S. Army Engineer Research and Development Center (ERDC) that led to the development of various mobility prediction methodologies including the NATO Reference Mobility Model (NRMM). These methodologies are numerical algorithms for predicting cross-country vehicle movement at length scales of several meters to several kilometers. They are based on empirical results drawn from years of resource-intensive experimental testing and have been used widely by the military community. As a consequence of their empirical nature, while the methods have been useful for prediction of large, heavy vehicle mobility, they can neither be extrapolated to today's vehicle designs containing advanced technologies nor are applicable to lightweight robotic vehicles. Hence is the need for a physics-based approach that would enable a high fidelity mobility prediction.

This research is aimed at augmenting the role that computational multibody dynamics (MBD) plays in characterizing the dynamics of discrete media in terramechanics applications. The discrete element method (DEM) is a powerful, emerging tool for analyzing these phenomena. The use of novel frictional-contact modeling techniques, such as the differential variational inequality (DVI) approach, which complement existing penalty/regularization approaches, in combination with a mapping of the solution algorithms onto commodity massively parallel hardware is envisioned to lead to a major breakthrough in our ability to simulate the dynamics of tracked and wheeled vehicles operating on deformable terrain.

The challenge, however, is the computational performance of such approaches in terms of efficiency and robustness which are addressed in this research. The research focuses on the use of second order optimization methods to improve the robustness of the DVI method for large multiscale MBD problems. This research thrust is predicated on the assertion that bringing higher order information into the numerical solution will improve its rate of convergence. Speaker(s): Paramsothy Jayakumar (US Army TARDEC)