PHYSICS-AWARE MODEL SIMPLIFICATION FOR INTERACTIVE VIRTUAL ENVIRONMENTS

Abstract

Rigid body simulation is an integral part of Virtual Environments (VE) for autonomous planning, training, and design tasks. The underlying physics-based simulation of VE must be accurate and computationally fast enough for the intended application, which unfortunately are conflicting requirements. Two ways to perform fast and high fidelity physics-based simulation are: (1) model simplification, and (2) parallel computation. Model simplification can be used to allow simulation at an interactive rate while introducing an acceptable level of error. Currently, manual model simplification is the most common way of performing simulation speedup but it is time consuming. Hence, in order to reduce the development time of VEs, automated model simplification is needed. The dissertation presents an automated model simplification approach based on geometric reasoning, spatial decomposition, and temporal coherence. Geometric reasoning is used to develop an accessibility based algorithm for removing portions of geometric models that do not play any role in rigid body to rigid body interaction simulation. Removing such inaccessible portions of the interacting rigid body models has no influence on the simulation accuracy but reduces computation time significantly. Spatial decomposition is used to develop a clustering algorithm that reduces the number of fluid pressure computations resulting in significant speedup of rigid body and fluid interaction simulation. Temporal coherence algorithm reuses the computed force values from rigid body to fluid interaction based on the coherence of fluid surrounding the rigid body. The simulations are further sped up by performing computing on graphics processing unit (GPU). The dissertation also presents the issues pertaining to the development of parallel algorithms for rigid body simulations both on multi-core processors and GPU. The developed algorithms have enabled real-time, high fidelity, six degrees of freedom, and time domain simulation of unmanned sea surface vehicles (USSV) and can be used for autonomous motion planning, tele-operation, and learning from demonstration applications.