基于位置推算算法的虚拟人感知机制的研究

Perception mechanism of virtual humans based on dead-reckoning algorithm

近十几年来，随着对行为动画研究的开展，研究者们越来越认识到行为研究对虚拟人研究的重要 性．感知机制是虚拟人行为研究的核心，迄令为止研究者们一般采用传感器方法而不是合成视觉方法建立 虚拟人的感知机制，这是因为在象虚拟战场这样的多虚拟人环境中，虚拟人本身的复杂性、环境的动态性，使 得虚拟人间可视信息的传递很困难．为了解决这个问题，提出了基于位置推算算法的虚拟人感知机制．该 项研究弥补了合成视觉方法的不足，丰富了行为动画中感知技术的研究，从而为虚拟人行为控制的研究奠定 了基础．

Presents the perception mechanism of virtual humans based on a dead-reckoning algorithm to overcome the drawback of the synthetic vision approach, promotes the study of perception techniques in behavioral animation, and lays the foundation for the behavior control study for virtual humans, and points out that with the development of behavioral animation in more than one decade, researchers have been more and more convinced of the importance of the behavior study for virtual humans; perception mechanism is the core of the a-hove study; the perception mechanism for virtual humans has been generally founded by the sensory approach, not by the synthetic vision approach because the complexity of virtual human itself and the dynamic quality of environments make it difficult to transmit visual information between virtual humans in multi virtual human environment like a virtual war field.