作者: Brian Vincent Mirtich
DOI:
关键词:
摘要: Dynamic simulation is a powerful application of today's computers, with uses in fields ranging from engineering to animation virtual reality. This thesis introduces new paradigm for dynamic simulation, called impulse-based simulation. The designed meet the twin goals physical accuracy and computational efficiency. Obtaining physically accurate results often whole reason performing however, many applications, efficiency equally important. Impulse-based simulate moderately complex systems at interactive speeds. To achieve this performance, certain restrictions are made on be simulated. strongest restriction that they comprise only rigid bodies. The hardest part body modeling interactions occur between bodies contact. most commonly used approaches penalty methods, followed by analytic methods. Both these constraint-based, meaning constraint forces contact points continually computed applied determine accelerations bodies. departure approaches, there no explicit constraints maintained points. Rather, all affected through collisions; rolling, sliding, resting, colliding modeled way. approach has several advantages, including simplicity, robustness, parallelizability, an ability efficiently classes difficult using constraint-based been experimentally tested sufficient applications. The processing collisions critical aspect approach. Efficient algorithms needed detecting large number occur, without missing any. Furthermore, simulator rests upon collision response algorithms. describes essential algorithms, their underlying theory. It how simple may extended articulated prove method truly practical, have implemented prototype simulator, Impulse. Many experiments performed Impulse described.