Spatial Reasoning for Machine Understanding of Solid Models

IRI-9414523亚利桑那州迈克尔·M·马雷法特大学，78,552-12个月。用于实体模型机器理解的空间推理这是一项为期三年的研究项目的第一年资助，其目标是研究和开发几何推理机制和知识表示方法，用于建模和识别将在工业设计和制造中有用的语义形状特征。为了以任何方式与物体互动，一个人或一台机器要感知物体，必须做的不仅仅是看到物体；有必要认识到它们是什么，赋予它们一个意义。实际上，将物体视为具有特定形状和特征的单独物品是感知的一部分，涉及几何推理。赋予对象意义意味着它们的重要之处必须有一个内在的知识表示。从一组几何特征到知识表示意义的映射称为语义，类似于语言中的映射。随着产品生命周期的缩短和产品种类的增加，最大限度地减少产品设计开始和生产时间之间的滞后期是至关重要的。该项目包括应用开发的几何推理模型和机制来解决机器对设计的理解、工艺规划策略的自动生成以及基于实体模型的自动视觉检查等重要问题。在这些问题上取得了初步的有希望的结果。将进行系统的进一步研究，以发展这些概念和机制。此外，还将研究这些概念和机制在开发面向柔性设计和生产的信息驱动智能集成环境中的应用。

IRI-9414523 Michael M. Marefat University of Arizona $78,552 - 12 mos. Spatial Reasoning for Machine Understanding of Solid Models This is the first year funding of a three-year research project with the objectives of studying and developing geometric reasoning mechanisms and knowledge representation methods for modeling and recognition of semantic shape features that will be useful in industrial design and manufacturing. A person or a machine that is going to perceive objects in order to interact with them in any way must do more than see the objects; it is necessary to recognize them as what they are, to attach a meaning to them. Actually seeing the objects as individual items with particular shapes and features is part of perception, and involves geometric reasoning. Attaching meaning to the objects means that there must be an internalknowledge representation for what is significant about them. A mapping from a set of geometric features to a meaning in terms of the knowledge representation is called semantics, after the analogous mapping in language. As the product life cycles become shorter and product variety increases, it is crucial to minimize the lag period between the start of a product's design and the time it is produced. This project includes applying the developed geometric reasoning models and mechanisms to solve important issues in machine understanding of designs, automated generation of process planning strategies, and automated visual inspection based on solid models. Initial promising results have been obtained on these problems. Further investigation will be performed systematically to develop the concepts and the mechanisms. In addition, the application of the concepts and mechanisms to the development of information driven intelligent integrated environments for flexible design and production will be studied.