基于模糊推理的人工鱼嗅觉感知研究

Research on Virtual Smelling Perception System of Artificial Fish Based on Fuzzy Reasoning

"人工鱼"利用人工生命方法实现计算机动画。为了丰富人工鱼的感知能力,提高对自然鱼的逼真动画效果,文章在分析"自然鱼"嗅觉机制、参考"机器鱼"感知系统的基础上,研究并实现了人工鱼的虚拟嗅觉感知。首先在人工鱼体系结构基础上建立了人工鱼嗅觉感知模型,并仿照生物嗅觉形成机理设计了三层嗅觉识别模型,采用基于主成分分析的模糊推理方法实现了嗅觉感知。经动画仿真,得到满意结果。嗅觉感知的研究和实现为人工鱼多感知融合系统提供了基础,也可为机器鱼嗅觉研究提供软件仿真平台。

Artificial fish for computer animation is created with artificial life. In order to enrich perceptive abilities of artificial fish and exhibit vivid animation, a kind of virtual smelling perception system of artificial fish was proposed based on analysis of the olfaction mechanism of natural fish and perception system of “robot fish”. The smelling perception model was built based on analyzing the frame system of artificial fish. Three-layer recognition model of smelling perception was designed according to the biologic mechanism of olfaction. The recognition of smelling information was achieved using fuzzy recognition method based on principal components analysis. The result is satisfied in computer animation simulation. And the design of smelling perception system sets a good foundation for research on multi-perception fusion and supplies a software simulation platform for the researching on recognition of smelling of robot fish.