ITR/ASE: 3D Sketching for Conceptual Visualization, Simulation, and Learning

ITR/ASE：用于概念可视化、模拟和学习的 3D 草图

• 批准号: 0428133
• 负责人: Hod Lipson
• 金额: --
• 依托单位: Cornell University
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2004
• 资助国家: 美国
• 起止时间: 2004-12-01 至 2010-11-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0428133&HistoricalAwards=false
• 关键词: ITR; ASE; 3D; Sketching; Conceptual

The enormous potential of sketch-based interfaces for conceptual engineering design is widely recognized and has been broadly pursued, but the practical application of these systems has remained limited because they are two-dimensional and do not embody functionality. The PI has previously demonstrated algorithms based on statistical machine learning that allow reconstruction of a 3D object from a 2D freehand drawing directly. By extending and combining this work with recent mass-market pen-based hardware, the PI will in this project create the basis of new conceptual design tools for rapid visualization, simulation and learning for geometry-intensive domains. A new form of freehand sketching will be explored, in which the sketch itself is spatially reconstructed in interactive time and then can be physically simulated, a process the PI calls "live sketching." The PI's system will enable users to construct complex 3D models quickly and intuitively, to analyze them, to change viewpoints while sketching, and to communicate and collaborate on spatial concepts with other designers more easily. To this end, the PI will first extend his reconstruction algorithms through context sensitivity and temporal scene segmentation so they can handle more complex geometries than is currently possible, including curved surfaces. This effort will then be combined with existing simulation and analysis codes of various disciplines and will be integrated into a prototypical platform allowing construction of 3D models by successive piece-wise sketching, reconstruction, and interactive simulation. The resulting tools will fundamentally alter the way engineers interact with design systems, bringing together traditional computer-aided engineering and simulation tools so they can play a role much earlier in the critical preliminary design stages. Broader Impacts: The outcomes of this project will positively impact any design domain where it is necessary to predict the behavior of spatial geometries, including many application areas of mechanical, civil, MEMS, and architectural CAD. Live sketching will also make engineering tools for modeling and simulation much more accessible to non-engineering designers, students and teachers, by removing interface barriers. The PI will in addition develop and make available online a live sketch application for teaching structure and kinematics, which will allow students and teachers to explore and demonstrate kinematics and structural analysis of trusses and solid structures by sketching structures, applying loads, and seeing the resulting mechanism trajectories, exaggerated deformations, and colored stress concentration for various shapes, in real time while sketching. This project will be made available on-line in open-source format, so that additional types of simulation can be integrated into it by others as required.

基于草图的界面在概念工程设计中的巨大潜力得到了广泛的认可，并得到了广泛的追求，但这些系统的实际应用仍然有限，因为它们是二维的，不体现功能。 PI先前已经展示了基于统计机器学习的算法，该算法允许直接从2D手绘图重建3D对象。 通过扩展这项工作并将其与最近的大众市场笔式硬件相结合，PI将在该项目中为几何密集型领域的快速可视化，模拟和学习创建新的概念设计工具的基础。 一种新形式的手绘草图将被探索，其中草图本身在交互时间内被空间重构，然后可以被物理模拟，PI称之为“现场草图”。“PI的系统将使用户能够快速直观地构建复杂的3D模型，分析它们，在绘制草图时改变视角，并更容易地与其他设计师就空间概念进行交流和合作。 为此，PI将首先通过上下文敏感性和时间场景分割扩展其重建算法，以便它们可以处理比目前可能的更复杂的几何形状，包括曲面。 这项工作将结合现有的模拟和分析代码的各个学科，并将被集成到一个原型平台，允许建设的3D模型，连续分段草图，重建和交互式模拟。 由此产生的工具将从根本上改变工程师与设计系统交互的方式，将传统的计算机辅助工程和仿真工具结合在一起，使它们能够在关键的初步设计阶段更早地发挥作用。 更广泛的影响：该项目的成果将对任何需要预测空间几何行为的设计领域产生积极影响，包括机械，土木，MEMS和建筑CAD的许多应用领域。 通过消除界面障碍，实时草图还将使非工程设计师、学生和教师更容易使用建模和模拟的工程工具。 PI还将开发并在线提供一个用于教学结构和运动学的实时草图应用程序，这将允许学生和教师通过绘制结构草图，施加载荷，并在绘制草图时真实的时间内查看产生的机制轨迹，夸张的变形和各种形状的彩色应力集中来探索和演示桁架和实体结构的运动学和结构分析。 这一项目将以开放源代码的形式在网上提供，以便其他方面可以根据需要将其他类型的模拟纳入其中。