Collision detection and motion generation for virtual manufacturing simulator

This thesis presents algorithms for exact and efficient modeling of collision detection and motion generation in VR manufacturing simulator. Collision detection and motion control are critical aspects of Virtual Reality (VR) based simulation of manufacturing processes. To achieve simulation of complex systems at interactive speeds most commonly used simulation packages and high performance graphics libraries trade accuracy for computational efficiency in modeling collision detection and motion generation. The research problem motivating this thesis has originated from the need of developing reliable, accurate VR manufacturing simulators which can be used not only as a visualization tool but also as a strategic validation tool.; Limited and slow progress has been reported in developing collision detection algorithms for general, nonconvex objects. We developed a technique which extends applicability of exact collision detection algorithms from convex to nonconvex objects. The key notion is the creation of Virtual Objects (VO) associated with specific scene objects and testing collision between VO instead of testing collision between the original scene objects. We introduced a methodology to encapsulate into VO the surface patches of interest for collision detection as well as the automatic procedure for creation of VO and for partitioning them into convex pieces. Motion generation tools presented in this thesis enable users to have full control of motion parameters during entire simulation. Instead of using computationally expensive numerical analyses routines, a very accurate local technique is used for estimation and derivation of motion parameters. In case the trajectory is not generated based on the laws of simulated physics, Bezier spline based path specification is employed. The algorithms described in this thesis are best suited for interactive simulation and animation applications where high accuracy of object modeling and motion control are essential. Examples include NC machining operations, robotic welding, robotic painting, virtual assembly and mobile robot simulation.