KDI: Virtual Environments to Elucidate Strategies in Complex Spatial Problem Solving

虽然空间认知已经在许多领域得到了研究，但是对于空间技能在解决复杂问题中的整合仍然知之甚少。 在这项研究计划中，虚拟环境技术（VET）将被用来研究复杂的空间技能，并从本研究的结果，开发有效的虚拟环境来训练这些技能。 凭借其交互性和引人注目的性质，虚拟环境可以成为培训，可视化，规划和沟通的基本工具，在无限的各种应用程序。 但是，如果VET是达到它的潜力，必须有一个更好的理解人类如何获取和整合在三维空间中的多感官信息。工作的重点应用将是外科手术培训，这为研究空间认知提供了一个具有挑战性的环境。外科手术任务的执行需要广泛的空间过程，以便使用复杂的空间表示来计划，导航和推理。 外科医生必须根据X射线、CT、MRI或超声图像的表面视图或横截面，形成三维解剖结构的心理图像。 根据这个模型和基于经验和解剖学知识的目标状态，他或她必须计划一个策略来获得重要解剖结构的暴露并获得期望的结果。 随着微创技术的出现，外科医生依赖于内部解剖结构的视频图像，并使用在穿过皮肤时由支点限制的器械。这项工作有两个协同的推力。 第一个项目旨在通过研究空间认知技能如何整合到复杂任务的执行中来推进认知科学的基础知识。具体来说，该项目将研究小规模三维环境中的导航，可变形结构之间机械相互作用的心理模拟，以及这些空间和机械推理过程如何整合到解决复杂空间问题中。 这项工作将指导计算模型的建设3-D空间模型从多个视图。 第二个重点是通过设计智能系统来训练和帮助人类解决空间问题，从而提高人机交互的艺术水平。

Although spatial cognition has been studied in a variety of domains,there is still little known about the integration of spatial skills insolving complex problems. In this research program, virtualenvironment technology (VET) will be used to study complex spatialskills and, from the results of this study, to develop effectivevirtual environments to train these skills. With their interactiveand compelling nature, virtual environments can become an essentialtool for training, visualization, planning, and communication in anunlimited variety of applications. But if VET is to reach itspotential, there must be a better understanding of how humans acquireand integrate multi-sensory information in three-dimensional space.The focus application of the work will be surgical training, whichprovides a challenging environment for studying spatial cognition.The performance of surgical tasks requires a broad range of spatialprocesses in order to plan, navigate, and reason using complexrepresentations of space. The surgeon must develop a mental image ofthree-dimensional anatomy based on a surface view or cross sectionsfrom X-ray, CT, MRI, or ultrasound images. From this model and a goalstate based on experience and anatomical knowledge, he or she mustplan a strategy to gain exposure of the important anatomy and obtainthe desired result. With the advent of minimally invasive techniques,the surgeon relies on a video image of the internal anatomy and useinstruments constrained by a fulcrum at their passage through theskin.This work has two synergistic thrusts. The first aims to advancefundamental knowledge in cognitive science by examining how spatialcognitive skills are integrated in the performance of complex tasks.Specifically, the project will study navigation in small-scalethree-dimensional environments, mental simulation of mechanicalinteractions between deformable structures, and how these spatial andmechanical reasoning processes are integrated in solving complexspatial problems. This work will be guided by computational models ofthe construction of 3-D spatial models from multiple views. Thesecond thrust is to advance the state of the art in human-computerinteraction by designing intelligent systems to train and assist humanperformance in spatial problem solving.