Research On Knowledge Base Management System For The Behavior Decision Of Unmanned Vehicle

Using traditional methods to build knowledge base can not meet the actual needs,especially in the field of unmanned driving.How to effectively acquire and represent knowledge from the behavior decision of unmanned vehicles,how to deal with the perceived vehicle information reasonably,so as to store,manage and use it,and build a knowledge base,how to solve and infer the behavior decision-making problem of unmanned vehicle accurately in complex scenes.These are the key problems to be solved in the knowledge base.In order to solve the key problems of knowledge acquisition and representation,knowledge modeling and fusion,knowledge reasoning and storage,this dissertation adopts the concept of knowledge base management system,and develops behavior decision-making knowledge representation,knowledge consistency test,problem solving,modeling method,multi knowledge fusion reasoning based on multi-level knowledge hypergraph,as well as for unmanned vehicles research on the construction method of vehicle behavior decision knowledge base management system.The automatic progress of user query condition semantics is realized by computer line reasoning.Through the construction and experiment of the knowledge management system of behavior decision-making of unmanned vehicles,the research results have strong pertinence and practicability.In the future,it has broad application background and huge application demand,and also has certain frontier in theoretical research.The main contents of this dissertation are as follows:1)Research on knowledge acquisition and representation of behavior decision making based on multi level knowledge hypergraphThe representation and fusion method of rules,cases,schemes and other knowledge needed in decision-making of unmanned vehicles are studied.Aiming at the problems that there is no uniform semantic description of the behavior decision information of unmanned vehicles,the required knowledge can not be quickly queried,and the semantic fusion of relevant knowledge is difficult,a framework for autonomous vehicle behavior decision-making in complex scenes is proposed.That is,the knowledge acquisition and representation method of autonomous vehicle behavior decision-making based on multi-level knowledge hypergraph.A knowledge consistency checking method based on multi-level knowledge hypergraph is designed and implemented.2)Research on knowledge modeling and reasoning method of behavior decision making based on multi level knowledge HypergraphThis dissertation analyzes the standardized representation of decision model,rule,case,scheme and so on,and constructs the behavior decision knowledge modeling method based on multi-level knowledge hypergraph.Thus,the visual acquisition,management,storage,maintenance and integration of decision-making knowledge of unmanned vehicles are realized.In addition,the multi knowledge integration,multi system collaboration,multi-level master-slave reasoning mechanism and multi knowledge fusion are analyzed,and the multi-knowledge reasoning fusion algorithm is designed to realize the knowledge service of multi-knowledge fusion and collaborative reasoning.3)Construction of behavior decision knowledge base management system for driverless vehiclesThe knowledge base management system of autonomous vehicle behavior decision based on multi-level knowledge hypergraph is constructed,including driving behavior knowledge base and traffic rule knowledge base,and the design of knowledge base management system is completed.A visual knowledge base management system is developed to realize the visual management,storage,maintenance and integration of decision-making knowledge of unmanned vehicles.In this dissertation,the knowledge acquisition,knowledge representation,knowledge modeling and reasoning methods,as well as the collaborative reasoning and multi knowledge fusion methods of multi-level knowledge hypergraph for autonomous vehicle behavior decision-making are realized,and the development of knowledge base management system for autonomous vehicle behavior decision-making is completed.