Research Of Key Technologies In Cloth Simulation Based On Physical Model

Since entering the information age, the computer animation technology has penetrated into people’s daily life bit by bit. Realistic 3D film effects, dazzling three-dimensional advertising, life-like characters, these are the product of the development of computer animation. Cloth simulation can effectively enhance the sense of reality of computer animation characters, and the real dynamic simulation effect of the curtains, tablecloths, flags and other cloth which have dynamic deformation because of the external force, friction force and other external influence, this makes the simulation effects of computer animation are more close to the real world. Since the cloth has the characteristics of motion and deformation with large displacement, large rotation and large strain, and the complex interactive environment of cloth, such as the effect of friction, resistance in the simulation process, these factors will affect the efficiency and performance of cloth simulation. However, it is difficult to apply all these factors to cloth simulation system for real-time, so how to provide cloth with the real-time, stabilized simulate is still a challenge for researchers.This paper discusses the key technology of cloth real-time animation based on the above background. Through researching the construction methods and numerical methods of cloth model, the comparative analysis of collision detection and response methods, this paper constructs the cloth mesh based on mass spring model and establishes the cloth simulation equation of Newtonian mechanics with the solution of Verlet numerical integration method. At last, it uses collision detection based on optimized AABB bounding box and response method to simulate the motion of the cloth. The concrete research contents are as follows:1. Improved "super-elastic deformation" phenomenon solutions. Cloth simulation based on Mass-spring Model due to the assumption that the cloth spring constant is linear, so when the cloth particles is applied by large tensile force, it will happen "super-elastic deformation" phenomenon.This paper analyzes the reason of "super-elastic deformation" phenomenon and describes the traditional solutions, the new method based on the stretch tensor of "super-elastic deformation" solutions is proposed.2. Numerical Method. Using Newton’s second law to determine the deformation and particle trajectories based on mass-spring cloth model. At each time step, through the use of numerical solution method to solve the first order ordinary differential equation, determining the new location of the particles. There are many ways of numerical solver; this article will focus on the explicit integration method, implicit integration methods, semi-implicit integration methods, and Verlet numerical integration methods. Comparing the efficiency and stability of this method, selecting Verlet numerical integration method as our method.3. Collision detection and response algorithms. Collision detection and response is the bottleneck to achieve real-time cloth simulation. Collision detection and processing time is about half of the total simulation time, so collision detection algorithm is very important. We use the optimizing AABB bounding box method in this paper as collision detection and response algorithm.This paper has two innovations as following:1. Proposing the based on stretch tensor particle sorting algorithm to solve particle-spring model "super elastic deformation" phenomenon. Stretch tensor can be characterized by the degree of particle stretch, so we first order the particles according to stretch tensor in descending order, Correcting larger stretch tensor particle position using Provot method, avoid cloth concussion problems caused by traditional methods in the correction process.2. Using regional pretreatment of AABB bounding box method, reduces the storage space, improve the efficiency of collision detection and processing, to enhance the simulation authenticity. Before the collision detection we pretreating the AABB tree for node storage optimized to reduce the memory, since self- simulation process, we excluding impossible self-collision detection area to reduce the workload of self-collision detection, improving real-time simulation.In summary, this paper based on computer graphics theory, Newton’s laws of mechanics and flexible object simulation, ensuring the cloth real-time simulation as well as improving the realistic of the cloth simulation, this method provides a methods and techniques about the solution of "super-elastic deformation" phenomenon and improvement of the efficiency of the cloth self-collision detection methods. The research has a very broad application space in 3D virtual fitting clothing, clothing production and presentation software, movies, games, animation and other fields.