Research On Virtual Intelligence Entity Based Artificial Intelligence

Virtual Reality(VR) technology is a very important branch among computer science. It would play important role in industry design & manufacture, training & education, military simulation, remote medical treatment & surgical operation, game, entertainment and so on. As a flouring and rapid developing high technology, virtual reality has already take great influence in human work and everyday life. Supported by the Guangdong province science center of the digital park D19--"Virtual ping pang", the work of this paper is to do research on artificial intelligence of virtual reality. The problems in this dissertation are that how the state of Virtual Intelligence Entity changes and how the action of Virtual Intelligence Entity do and path planning in the virtual environment. And the new ideas and the new methods are proved in this dissertation for developing a new system of virtual reality and studying. There are some creative fruits which achieved in it.In this dissertation the background of virtual reality technology, the related theories of artificial intelligence in virtual reality domain are firstly summarized.The method basing probability deduction for deciding the action of virtual intelligence entity is gained in the second part. This is feasible. And this can also improve the intelligence level of virtual intelligence entity. There are two ways: one is Bayesian theory and the other is Dempster-Shafer theory. At the end of this part the differences are embodied.Then aiming at the three I which is Immersion,Interaction,Imagination in virtual reality technology, in this dissertation put forward a coordinate model of 6 degrees of freedom. The most important ingredient in the three I is interaction. Interaction is the cause for the successful systems of virtual reality. The model in this dissertation can solve the problem of the states changing of virtual intelligence entity. This can make virtual intelligence entity real time follow the user. At the end of this part, one typical example is used to prove the effect of this means.In addition, there are some traditional methods for path planning, such as potential field method and probabilistic road map method. Those methods are just only applied in Robots, which are the same with static obstacle. In this dissertation, there is a new path planning method. Make use of mesh partition. Then use this method in virtual environment. The complexity of this method is O ( n log n ).With some tests, the method has preferably efficiency. The most important is that this method can use for dynamic Obstacle in a virtual environment. And the method can solve the difficult problems in virtual reality, such as path planning of camera, browsing in a complex virtual scene and how the virtual intelligence entity moves in a system of virtual reality.Finally, this thesis makes a summarization of the entire work and points out the future research direction.