生物领域知识激励的功能知识建模与类推方法研究

Traditional Biomimetic design realizes give designers enlightens about how to promote new inventions and techniques in engineering through simulating the biological functions, working principles, strategies and structures. However, biomimetic design focuses on designers’ own knowledge reuse. Lack of a systematic biological knowledge modeling and analogy approach leads to randomicity and contingency of biomimetic design. Thus, research on how to bridging mechanical design and biological knowledge and providing the process of simulating existing biological cases with great facilitates is necessary. Grounded on the characteristics of biological and mechanical design knowledge, this program proposed the biologically inspired design methodology, which classifies biomimetic design into: problem-driven and solution-driven biologically inspired design according to the start point of design. .Major goals of this research are as follows:.1. Propose a functional ontology based Constrained Function-Behavior Structure (CFBS) knowledge cell model for consistent knowledge modeling. The model is based on analyzing the knowledge obtained in multi-domain. The mapping relationships intra and inter CFBS knowledge cells are constructed. Cross domains function basis, which serves as the bridge between biological and engineering domain, are summarized and listed in appendix I and II,..2. Propose a CFBS knowledge cell clustering and retrieving approach. A combined sematic similarity and a hybrid numeric similarity calculation algorithm are proposed to comparing functional, behavioral and structural relevancy with the design objectives. CFBS knowledge cell clustering method works as a tool to obtain useful knowledge cells to designers..4. Propose the CFBS knowledge cell based biological domain knowledge reuse approach. Solution driven design model and problem driven bio-inspired design model are given separately. The models take across domain functional ontology as bridge and CFBS knowledge cells library as the design depository, employ knowledge cells clustering approach and design synthesis method as tools, make the cross domains retrieval, reuse possible..5. Develop prototype system and study design case. A prototype system, named BIDS (Biologically Inspired Design System), based on the proposed methodology is developed. Design of the implantable visual prosthesis is demonstrated to testify the effectiveness of the proposed methodology.

传统的仿生设计通过简单模仿生物领域存在的生物功能、行为原理和结构以启发工程领域设计活动。其将设计活动局限在从已知生物实例出发的对生物形态、现象的简单模仿，而忽视了对如何系统地利用生物领域知识激励工程领域设计活动的方法研究，导致设计结果存在极大的随机性和偶然性。.本课题意在从设计方法学的角度出发，对仿生设计进行深入讨论和研究，将知识工程方法引入仿生设计中并提出系统地生物激励设计方法，其包括：基于功能本体的生物和工程领域一致性知识表示模型以解决跨领域知识表示存在的不一致性；研究跨领域知识聚类检索方法以实现对已有生物领域实例知识的智能检索重用；根据设计活动的出发点不同，分别研究工程领域问题需求出发的问题驱动生物激励设计和生物领域已知实例出发的解驱动生物激励设计所对应的跨领域类推过程模型，并完成人机交互式生物激励设计系统的开发，以实现生物领域知识对工程领域创新设计活动的系统支持。

现有的仿生设计通过模仿生物领域存在的生物功能、行为原理和结构以启发工程领域设计活动。其将设计活动局限在从已知生物实例出发的对生物形态、现象的简单模仿，而忽视了对如何系统地利用生物领域知识激励工程领域设计活动的方法研究，导致设计结果存在极大的随机性和偶然性。本课题对仿生设计进行了系统地讨论和研究，提出了生物激励设计方法，根据设计活动的出发点不同，将生物激励设计划分为：从工程领域问题需求出发的问题驱动生物激励设计和从生物领域已知实例出发的解驱动生物激励设计。基于此完成构建了支持跨领域知识重用的一致性知识表示模型，研究了跨领域知识检索算法并提出了生物激励设计解方案生成与评价方法，实现了生物领域知识激励工程领域设计活动。. 本课题通过研究生物领域和工程领域知识的特点，采用知识工程的相关技术，从功能的角度出发，提出了基于功能本体生物知识建模与重建及应用方法以实现生物领域知识对工程领域设计活动的跨领域支持。本课题主要的研究成果如下：. 1. 提出了基于功能本体的生物和工程领域跨领域一致性知识表示模型。. 2. 基于生物领域和工程领域知识元模型，提出了融合功能语义相关性和行为特征参数相合性地跨领域知识聚类检索方法。. 3. 提出了问题驱动和解驱动的生物领域知识的跨领域重用过程模型。. 最后，基于以上方法开发了跨领域知识元建模模块和生物领域知识在工程领域重用过程模块构成的人机交互式生物激励设计系统（BIDS，Biologically Inspired Design System）。. 综合上述成果，本研究为工程领域创新设计奠定了坚实的工作基础。同时为设计活动开展提供了系统的方法支持。项目资助发表SCI论文7篇，EI及核心论文5篇，国际会议并做大会发言1篇。培养博士生4名，其中2名已取得博士学位.项目投入经费20万元，各项支出基本与预算相符，剩余经费计划用于本课题研究后续支出。