規則式推論情境感知系統之設計

Abstract

目前智慧型手機的發展蓬勃，普及率也日漸提高，不論是硬體或是軟體技術都越來越進步。目前手持裝置上有許多感應器可以取得使用者周遭的環境資訊，再加上wifi、3G等連線功能，使得我們可以發展有如隨身行動助理般的情境感知(Context Aware)功能。情境感知的最主要目的是讓手持裝置更加的智慧化，盡量的減少使用者的輸入，主動提供使用者最需要的資訊、服務，以解決手持裝置上輸入不易、顯示螢幕尺寸小帶來的不便。 想要完成一個情境感知系統必須從兩大部分著手，一部分是情境資訊的取得，另一部分為情境推理的過程。本研究以本體論(Ontology)描述環境，以規則為基礎進行推論來實現情境推理。本研究根據各種不同場景(scene)以及模式(mode)下建立相關的Ontology來描述環境，再以不同的參數、SWRL規則來進行推理，藉由推理出的結果可以得到使用者可能的情境或是直接回傳使用者可能需要的服務。 本體論可以明確的表達知識以及詳細的描述環境，如此一來我們可以明確的定義我們想要完成情境感知服務的各種場景。除此之外我們選用SWRL作為規則語言來直接對本體論下的實體撰寫規則，使得整體的推論完整，而本體論以及SWRL都是以OWL語言表示，在進行推理時不需要做任何語言轉換，整個推理過程顯得更有效率。本研究也提供使用者操作介面，使用者可以在介面上觀看情境資訊、服務資訊或是操作代表使用者位置的圖示，如此一來可以使用者可以更深刻的感受到情境感知服務帶來的便利以及魅力。Recently the trend of smart phone is booming and its penetration rate is rising. The smart phone related technologies are progressing very fast. Currently, context aware services based on hand-held sensor devices which can obtain the user’s context information are getting popular and are used as a mobile assistant. The main purpose of context awareness is to make handheld devices be more intelligent, which includes reducing the need of user input as much as possible and providing users the most needed information and services. The development of a context aware system includes two parts: first is to obtain the context information from sensors, second is to execute the context reasoning process. In this study, we use ontology to describe the environment according to different scenes and modes, in addition, rule-based reasoning based on SWRL rules is established to achieve context reasoning. Finally, the reasoning results derived from the SWRL rule engine can be used to call the predefined services in different scenario. In our system, both ontology and SWRL rules are based on OWL language. We also develop a graphic user interface to provide the user to monitor the progress of context scenario including the context information, service information, or action of the user.