摘要
针对目前人工智能技术在造型知识表达上的不足 ,提出了一种面向造型设计的知识表示法。以分层约束满足为基础 ,用多类谓词、拓扑限定、不确定性描述等方法使复杂的造型设计对象和过程得以清晰描述。通过将复杂的空间模型以结构分解和拓扑分解的方法离散为线性域 ,使用空间对象的变形映射为点的线性移动 ,从而使正向的造型过程和逆向的回溯过程都能沿可控路径自动进行。分析了设计知识表达的关键点 ,阐述了面向造型设计的知识表示法细节 ,结合一个分模设计的实例分析了空间推理过程 。
The representation of design knowledge is as the first important task for the intelligent modeling design. It influences directly the complexity, algorithm and automation of the design process. An effective representation of such knowledge must describe the modeling process explicitly, reflecting expert experiences and suitable to spatial reasoning. This paper proposes a modeling-oriented knowledge representation to make up the shortage of current methods on these aspects. The method is based on layered constraints, multi-types predications, topological restriction and uncertain description,itcan dissolve the deformation of the complex 3D models to the moving in linear field, therefore enable controllable paths for forward modeling and backward retracing. The paper analyses key points of the design knowledge representation, elaborates the modeling-oriented knowledge representation, and illustrates the spatial reasoning process with an example of mould-parting design. As proved by the results,proposed method can reduce the manual interference in the complex modeling design, shorten the design period and bring optimized results by experience inspiration.
出处
《中国机械工程》
EI
CAS
CSCD
北大核心
2001年第6期677-682,共6页
China Mechanical Engineering
基金
国家自然科学基金资助项目 (5 0 0 75 0 71
关键词
造型设计
知识表示
空间推理
拓扑限定
人工智能
modeling design knowledge representation spatial reasoning topological restriction
