Study And Implement Of Group Animation Based On Autonomous Agents

In computer group animation creation technology, the artificial life method of computer animation has radically overcome the defects of traditional animation creation technology and greatly improved animation creation efficiency.However the animation role's model manufacturing technology that this method adopts is becoming more and more complicated and the coupling degree between each model of animation system improves day by day too, which cause the difficulty of animation creation to strengthen day by day, Especially when increasing in intellectual role's figure, the calculation amount will increase so non-linearly fast that it has greatly influenced the animation real-time character and limited the wide application of this method.Therefore, based on the premise that animation effect and real-time character are not reduced, how to reduce the coupling degree between each model of animation system, decrease the calculation amount of the computer and satisfy the animation creation requirement of real time has important meaning and value for large-scale group animation creation. Aiming at this problem, from the design and realization technology of animation role model both, this paper has carried out thorough research and implementation.The main job and creative points in this dissertation are as the following:1) The dissertation analyzes and designs the animation role's model. Mainly aiming at these existing problems of animation role model coupling degree high and design difficulty big, we have designed a two-layer decomposition group animation role model based on autonomous intelligent agent. This model can reduce the model coupling degree, decrease the calculating amount and effectively satisfy the real-time character demand of large-scale group animation creation by resolving the role's sport animation into two parts of roles' behavior movement and behavior action, and each part adopting different implementation techniques.2) The dissertation has designed "space_separated" perception algorithm. Simulating the perception ability of role is one of the performance bottlenecks of large-scale group animation creation. In order to improve the role's perception performance, we have utilized the limited perception property of role and designed one "space_separated" perception algorithm. This algorithm has effectively reduced the potential perception entity number of role, improved the perception performance of role in complex virtual environment and strongly guaranteed thereal-time character of animation creation by separating three dimensional spaces.3) The dissertation has thoroughly studied the object-oriented implementation method of finite state machine. The finite state machine is a kind of effective behavior modeling method and can well portray role's behavior decision-making in the complicated virtual environment. In view of the implement shortcoming of the finite state machine based on traditional structural method, combining the state design pattern, we design and study one kind of object-oriented implementation method of finite state machine to overcome the shortcoming that the traditional structurization method exists.4) The dissertation adopts Steering behavior theory to portray behavior movement of the animation role. Mainly aiming at the existing problems of animation role movement model design and implementation difficulty, we decrease the design and implementation difficulty of the animation role's biomechanics model, reduce the calculation amount and further guarantee the real-time character of large-scale group animation creation by useing Steering behavior theory to portray behavior movement of the animation role.