CRII: CHS: Perceptual Data-Guided Computational Design

感知决定了人类如何解释和使用设计，因此是设计的重要考虑因素。这个CRII项目研究如何感知数据可以用来指导计算设计过程。作为展示，PI将研究如何从虚拟环境中获得的视觉感知数据可以用来设计新的计算方法，能够分析，可视化和建模给定的建筑布局的人类启示方面。这种方法可以应用于为各种建筑环境生成功能布局设计，以提高运动舒适度、效率、生产力和安全性。拟议的研究将对改善交通、旅游、商业和公共安全产生广泛影响。此外，PI将通过传播收集的视觉感知数据集做出贡献;定期组织视觉和认知研讨会;指导一个本科生，硕士和博士生团队进行研究，包括PI将通过McNair学者计划指导的少数民族学生。特别是，该研究所将建立一个配备眼动追踪功能的虚拟现实平台，用于研究人类对不同几何形状的视觉感知和反应，3D环境中的空间和语义特征。收集的数据将成为开发3D场景分析工具的基础，设计师可以使用该工具来可视化和了解布局设计对人类视觉感知和人流的影响，从而做出明智的设计选择。设计师还可以通过由优化组件驱动的建议性界面轻松获得人体工程学设计建议，该优化组件考虑了多个视觉感知和人类启示因素。 基于视觉感知数据，还将设计能够自动生成和增强3D布局设计的优化方法。

Perception determines how humans interpret and make use of designs, and is therefore an important consideration for design. This CRII project investigates how perceptual data can be used to guide computational design processes. As a showcase, the PI will investigate how visual perception data obtained from virtual environments can be used to devise novel computational approaches capable of analyzing, visualizing and modeling the human affordance aspects of given architectural layouts. Such approaches can be applied for generating functional layout designs for a variety of built environments, to enhance movement comfort, efficiency, productivity and safety. The proposed research will bring broad impacts on improving transportation, tourism, commerce and public safety. In addition, the PI will contribute by disseminating the collected visual perception datasets; organizing vision and cognition workshops regularly; mentoring a team of undergraduate, master and PhD students to conduct research, including minority students whom the PI will mentor via the McNair Scholars Program.In particular, the PI will set up a virtual reality platform equipped with eyetracking capabilities for studying human visual perception and reactions towards different geometric, spatial and semantic features in 3D environments. The collected data will form the basis for developing a 3D scene analysis tool which designers can use to visualize and understand the effects of layout designs on human visual perception and people flow, to make informed design choices. Designers can also easily obtain ergonomic design suggestions via suggestive interfaces driven by an optimization component which considers multiple visual perception and human affordance factors. Based on the visual perception data, an optimization approach will also be devised capable of automatically generating and enhancing 3D layout designs.