학술논문
시멘틱 추론기반 상황인식 프레임워크의 구현 / An Implementation of A Context-Aware Framework based on Semantic Reasoning
Document Type
Dissertation/ Thesis
Source
Subject
Language
Korean
Abstract
일반적으로 상황인식 서비스는 사용자의 다양한 의도와 상황을 이해하고 많은 데이터를 처리할 수 있는 상황인식 시스템을 필요로 한다. 상황인식 시스템이란 외부 사용자의 간섭 없이 현재 상황에 스스로 자신의 오퍼레이션을 적응하는 시스템이다. 이러한 상황인식 시스템의 개발을 수월하게 하려면 상황인식 프레임워크를 필요로 한다.본 논문은 SWRL 규칙을 포함한 OWL 온톨로지로 표현되는 시멘틱 추론에 기반한 상황인식 프레임워크를 제안하고 있다. 제안된 상황인식 프레임워크에서는 지식베이스에 저장된 모델(model), 사실(fact), 규칙(rule)이 모두 OWL 구문으로 표현되어 지식베이스 구조가 간결하며, 규칙 표현에 모델과 사실 표현을 그대로 사용할 수 있다. 그리고 이동 에이전트 기술의 사용으로 상황인식 서버와 클라이언트 애플리케이션을 네트워크 상에 수월하게 분산할 수 있고, 다수의 분산된 클라이언트들을 둘 수 있어서 확장성 요구사항을 해결하고 있다.그러나 현재까지 공개된 SWRL 추론엔진이 존재하지 않아서, 추론시점에 지식베이스에 있는 모델, 사실, 규칙들을 외부 규칙엔진에 보내어 추론하게 한 후 결과를 받아 지식베이스를 갱신하는 구현방법을 현재 사용하고 있다. 그 결과 상황인식 서버가 실행시간의 상당 부분을 상황 추론에 소요하여 응답지연의 결과를 초래하고 있다. 향후 시멘틱 추론을 효과적으로 처리하여 성능을 개선하는 후속 연구가 있어야 할 것이다.
The context-aware service is requiring context-aware system which understands user's multi-purpose and context and deals with more data. Context-aware system is able to adapt their operations to the current context without explicit user intervention. To facilitate the development of context-aware system a context-aware framework is needed.This thesis is proposing the context-aware framework based on semantic reasoning which inferences the context represented as OWL ontologies and SWRL rules. The proposed context-aware framework provides concise structure of knowledge base and makes the rule representation to use OWL model and facts because it represents model, facts and rules of the context with OWL syntax. Also with using mobile agent technology that can easily distributes context-aware context server and client applications, it resolved the requirements for both multi-distributed clients and scalability.Until now SWRL reasoning engines haven't developed. Hence the propose framework adapted the method that sends facts and rules to the external rule engine and then updates the knowledge base with the reasoning results after rule-based reasoning. This method has shown the probn t that most of context server's run-time is the time of the context reasoning and results in the response delay. In the future it is required that further researches improve the performance of the semantic reasoning.
The context-aware service is requiring context-aware system which understands user's multi-purpose and context and deals with more data. Context-aware system is able to adapt their operations to the current context without explicit user intervention. To facilitate the development of context-aware system a context-aware framework is needed.This thesis is proposing the context-aware framework based on semantic reasoning which inferences the context represented as OWL ontologies and SWRL rules. The proposed context-aware framework provides concise structure of knowledge base and makes the rule representation to use OWL model and facts because it represents model, facts and rules of the context with OWL syntax. Also with using mobile agent technology that can easily distributes context-aware context server and client applications, it resolved the requirements for both multi-distributed clients and scalability.Until now SWRL reasoning engines haven't developed. Hence the propose framework adapted the method that sends facts and rules to the external rule engine and then updates the knowledge base with the reasoning results after rule-based reasoning. This method has shown the probn t that most of context server's run-time is the time of the context reasoning and results in the response delay. In the future it is required that further researches improve the performance of the semantic reasoning.