Ontology-based Portable Campus Guidance System

本論文提出以能提供使用者可適性遊覽，並且能察覺作用中的環境參數的校園導覽系統。校園導覽的需求和導航系統與博物館數位導覽系統不同。校園導覽應該要提供環境中使用者有興趣的景點資訊，但是導航系統只能引領使用者一條到目的地最短的路徑，並且呈現出沿途景點的位置。這些資訊包含標的物相關的知識與校園導覽員的經驗。另外，校園導覽還應該要整理這些景點間的相互關係，可是博物館數位導覽系統卻只會介紹在使用者面前的展覽品。於本論文中，校園導覽應該有彈性去符合參觀者多樣的要求，而這樣的系統應該利用使用者的個人資料和環境中的資訊，來可適性地導覽參觀者。套系統包含兩個主要的技術：環境認知與知識的具體化。環境認知在處理一些環境中相關的變化，而知識的具體化在使用本體論(Ontology)在特定領域表達一些概念的集合，以及它們之間的關係。本系統中，環境認知的目的在處理使用者的概況、興趣、所在地，及時間的資訊。本系統從輸入獲得參觀者的資料，和從GPS接收器取得使用者的位置資訊。導覽知識是根據利用本體論所建構以節點為主的概念，使具體化的功能得以從初步資料推論高階的資訊。以本體論為基礎的功能可以支援即時性的推論，以符合實際使用的情況。本論文整合了之前提及的本體論與環境認知技術，並且透過複數代理人平台ADE，成為一套系統。因此，此系統可以被移植到有Java平台的任何個人電腦，也為了台大校園導覽而實作一套可攜式系統，提供可適性的資訊。此外，校園導覽系統還可以協助訓練校園導覽員。最後，展示三個實用的範例，來證明此系統的能力。The thesis proposes a campus guidance system that is aware of the operating context to provide users adaptive tours. The requirements of the system are different from the use of the navigation system and the museum digital guide system. The campus guidance system provides the information of the scenic spots in the environment that users are interested in, but the navigation system only leads in a shortest path to the destination and displays the spots along the way. The spot information includes the knowledge of the campusuidance and the experience from the campus guides. Moreover, the campus guidance system also organizes the relationships between those spots, but the museum digital guide system just introduce the exhibits to users who are in front of them. In this thesis, the campus guidance system should be flexible according to various requirements of visitors, and this system should utilize data of visitors and information about the circumstances to guide themdaptively. he campus guidance system uses two main technologies: context-awareness and knowledge conceptualization. Context awareness deals with related changes in the environment. Knowledge conceptualization uses ontology to represent a set of concepts within a knowledge domain and the relationships between these concepts. In the system, the goal of context-awareness is to obtain and cope with the data from visitors'' profiles, interests and positions, and information of date and time. The system obtains visitor''s data by inputs and uses GPS receiver to get the visitor''s position. The campus guidance knowledge, based on the node-oriented concepts built by ontology, could infer high-level information with the context-aware data. The ontology-based knowledge could support real-time inference to meet the actual situations of usage. The thesis integrates these mentioned ontology and context-aware technologies into the system on a multi agent platform, JADE. Therefore, this system could be ported to any PC-based computer withava Runtime Environment. A portable system is implemented to provide adaptive information for tours of the NTU campus. In addition, the campus guidance system could assist the training of the campus guides. Three scenarios, normal guidance, overtime guidance and near lunchtime guidance, are demonstrated to verify the capabilities of the system.Chapter 1 Introduction 1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Problem Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5hapter 2 Background 7.1 Related Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . 7.1.1 Location-based system . . . . . . . . . . . . . . . . . . . . . 8.1.2 Ontology . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1.3 Guidance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.2 Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.2.1 Cyberguide . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.2.2 GUIDE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.2.3 HIPPIE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.2.4 LAMP3D and seeVT . . . . . . . . . . . . . . . . . . . . . . 17.2.5 Ontology-based systems . . . . . . . . . . . . . . . . . . . . 18hapter 3 Ontology-based Graph Representation 19.1 Ontology-Geographic Node (OGN) . . . . . . . . . . . . . . . . . . 19.2 Ontology-Geographic Region (OGR) . . . . . . . . . . . . . . . . . . 22hapter 4 Design and Implementation 25.1 Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.2 Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27.2.1 Ontology knowledge . . . . . . . . . . . . . . . . . . . . . . 28.2.2 Multi agent platform . . . . . . . . . . . . . . . . . . . . . . 33.2.3 Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34hapter 5 Experiments 39.1 System Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.2 Scenarios for Existing System . . . . . . . . . . . . . . . . . . . . . 40.2.1 Normal guidance . . . . . . . . . . . . . . . . . . . . . . . . 41.2.2 Overtime guidance . . . . . . . . . . . . . . . . . . . . . . . 42.2.3 Near lunchtime guidance . . . . . . . . . . . . . . . . . . . . 43.3 Estimation of Performance . . . . . . . . . . . . . . . . . . . . . . . 43hapter 6 Conclusion 47ibliography 4