Research And Implement On The Behaviors Of Multiple Intelligent Virtual Human Based On The Situation Calculus

Virtual Human is an important part of VR (Virtual Reality). Presently, it is an important content of VR study to construct the intelligent virtual human with autonomic behaviors by introducing AI (Artificial Intelligence) to VR. This kind of intelligent virtual human can interact with users and can design their behaviors dynamically, indicating some intelligence. Now, the intelligent virtual human with both ability of cooperation and autonomic decision-making of actions in dynamic environment is a new study hotspot.It is the aim that to implement the multiple intelligent virtual human with ability of cooperation in VRML by studying the cooperation behaviors of multiple intelligent virtual. The Situation Calculus and ConGolog are adopted as the foundation of aim-oriented behaviors reasoning autonomously in dynamic environment. So, the virtual human can reason autonomously according the aim situation in current virtual scene and plan a series of actions to achieve the aim situation, which will display in virtual scene by virtual human animation language. The main research contents in this paper are as follows:(1) The virtual human modeling in VRML. A virtual human modeling is completed according the H-Anim virtual human standard, which offers an entity for implementing the actions. At the same time, a formal representation of autonomic cooperation behaviors of multiple intelligent virtual human with dynamic environment in virtual scene is given.(2) The ConGolog is extended. Based on the agent communication language ACL, a communication protocol SCMP is proposed. The communication action in ConGolog talk is defined explicitly, based on which, a kind of Requirement/Service cooperation model is proposed. So, multiple intelligent virtual human can complete the aim-oriented cooperation behaviors plan autonomously in dynamic environment. Also, the true concurrency is imported to the ConGolog, solving the contradiction between the interleaved concurrency of ConGolog and the true concurrency of multiple virtual human actions in the virtual scene.(3) A virtual human animation description language VHADL is defined. Many kinds of actions of virtual human in the virtual scene can be described very well by using VHADL. The primitive actions in ConGolog can also be described further by VHADL. So, the plan in AI and the three-dimensional animation in VR can be connected dynamically through VHADL. At the same time, the problem of information absence between primitive actions in ConGolog and actions animation can be solved very well.(4) The research contents are simulated. The simulation of sending coffee demonstrates that the methods mentioned above are correct and effective in the process of implementing the cooperation behaviors of multiple intelligent virtual human in virtual scene.