Research On The Mechanism Of Cloth Simulation And Integrated Modeling

Dynamic cloth simulation is a typical example of simulation for flexible objects with large deformation. It is a difficult research topic, involving theories and technologies on realistic display in computer graphics and analysis on complex physical and mechanical properties of the fabrics. However, it has a promising prospect in visual animation, garment CAD, virtual fitting and virtual prototype of fabric-mechanical coupling system. Thus, it has paramount significance both theoretically and practically to study the dynamic cloth simulation.In this dissertation, several cloth-simulation related problems are well studied and the simulation realism and the simulation time are taken as the main indicator. Our research is based on the phenotype of a large number of simulation models and the simulation results, interpretation of the mechanism and nature of the structure models. By analyzing the influencing factors of physical parameters and forces on the motion patterns of mass points, we find it correlate with simulation results, which reveal the law of modeling; help established a novel objective method to evaluate the performance of various cloth simulation models. By studying the model law of nature as a start point, we find the method and conditions which can balance the simulation speed and authenticity. By expanding the breadth of the cloth simulation between flexible and rigid nature attributes, we achieve unified rigid-flexible modeling. The contributions can be concluded as follows:Firstly, by studying various models of cloth simulation, we clarify the phenotype of the model. The abstract from the scientific aspects of the model makes the structure and mechanism of the model more clearly. Proposed cloth simulation model is based on fabric geometry and physical properties. Their interaction in simulation shows that the physics attribute take dominate. It suggests that the cloth simulation research should focus on improving the physical properties.The deep analysis on mass-spring model simulation mechanism clarifies the mass point motion pattern under the law of force, and shows that the damping coefficient plays a critical role in the simulation; By establishing the correspondence between simulation results and the mass point trajectory, an objective method to evaluate the performance of cloth simulation is given, which will be useful in choosing optimal model parameters during simulation.By means of the theory used to characterize the relationship between the force and the movement of a mass point in dynamics and kinematics, a novel dynamic bidirectional iteration algorithm based on generalized inverse kinematics is proposed. It can improve the speed and the reality during simulation. The existence of the inverse motion in the mass-spring model is expounded. By setting up differential equations for inverse motion kinematics, we design the two-way motion synthesis processing algorithms, and give out initial position setting method for inverse kinematics. It is verified theoretically and experimentally that our bidirectional iterative algorithm is able to shorten the simulation time and improve the simulation reality.Under the direction of "combining the rigid and the soft ", an integrated dynamic simulation model is proposed to represent cloth on the platform which is widely used for the rigid simulation. Firstly, the necessity of unified modeling the cloth- machinery system is analyzed. Then unified modeling methods by means of and "combining the rigid and the soft" are proposed. Experimental results corroborates that the proposed model can exchange information with the mechanical system, which indicates that the goal of coupling the cloth with the machinery has been achieved elementarily. It is rigid as well as soft, more adaptable and easier to extend.Finally, the dissertation is summarized, and the shortcomings are also discussed. The direction of further study is pointed out. We state that another key problem, existed in simulating cloth or other soft bodies with large deformations, is how to improve the efficiency when solving differential equations. To this end, we propose to make use of the scheme of non-explicit solutions (also called as solution sequence) to solve the equations and represent the obtained solutions. In this way, the goal to improve the simulation speed without the loss of the simulation quality can be achieved.