Research On Method Of Patent Knowledge Acquisition And Its Supporting To Conceptual Innovation Design

With the era of knowledge economy coming,patent is a crucial knowledge resource to promote enterprise development.The increasing market challenge urgently demands methods,techniques and tools for patent knowledge acquisition with high efficiency to directly serve product innovation.After comprehensively analyzing the history and the current situation of conceptual design and patent research,aiming at product conceptual design,research on the problems of current computer-aided innovation was carded out. The main research and primary contributions are reflected in the following five aspects:Firstly,research on the models of product patent classification and patent principle knowledge representation was constructed.A method of patent document classification guided by design objective was proposed to make full use of patent resources for product conceptual innovation.The knowledge system for product innovation,consisting of innovation level knowledge,association knowledge,principle knowledge and patent language knowledge,were set up.The relationship between patent knowledge space and the existing design knowledge space was described on the basis of hierarchy theory of knowledge space in conceptual design.Adopting frame structure and concept graph,the representation model for multi-level patent complex-function principle knowledge was established,which provided sufficient reference to product principle knowledge automatic acquisition from patent.Secondly,research on method of working structure information extraction based on natural language understanding was put forward.By adopting non-deterministic finite state machine,the knowledge base for technical component knowledge extraction was set up.Semantic type of technical components was identified by special term dictionary.By taking frame semantics as theory foundation,patent verb semantic frame was constructed for identification of technical relations and their relating semantic roles.And the natural language understanding process of working structure information extraction was also presented combining with the key techniques of syntactic and semantic analysis,which were illustrated by USA patent.The Results showed that the method could enhance the capability of patent working structure information extraction and realize deepening application of patents in conceptual design. Thirdly,research on method of patent deeper principle knowledge explanation was done.The learning frame was constructed with design objective,principle relative knowledge,operation ruler and explanation learning algorithm.Effect feature model of domain knowledge was studied to support explanation learning and relating principle among function,effect,working structure were set up.The incomplete matching method based on concept graph was proposed to explain product work structure information from bottom to top to obtain product function and effect knowledge.The results showed that this method could promote the construction of product patent knowledge-base and fully reflect product design principle,and improve the enlightening function of patent to product innovation.Fourthly,research on method of conceptual innovation supported by domain patent knowledge was carded out.The principle and criteria for the judgment of product creativity was analyzed and the role of domain patents in product innovation was emphasized.Then a feature- objective-oriented general model for conceptual innovation design was proposed,which described the multi-path innovation design process.Based on hierarchy theory of knowledge space in conceptual design,the innovation solution strategies inspired by patents from the same domain were discussed and specific implementation of effect design combined with morphology analysis was explained. Finally practicality of this method was demonstrated by a successful instance of drill chuck products,which further validated its ability to rapidly grasp innovation opportunity and effectively increase innovation ability in conceptual design.Finally,research on realization and application of computer-aided innovation system supported by domain patent knowledge was put forward.Based on the theoretical basis of the above,a computer-aided innovation prototype system supported by patent knowledge was developed.The software systems comprised three main modules,namely patent knowledge acquisition,innovation design and knowledge management. Combining specific application in enterprises,several innovative products of drill chuck was developed and one of them has been granted national invention patent,and two are under evaluation as utility model patent.