透過您的圖書館登入
IP:18.191.180.252
  • 學位論文

情境式系統配置代理人於M2M之應用

Scenario-based Configuration Agents for M2M Applications

指導教授 : 許永真

摘要


Machine-to-Machine (M2M) 是近年來興起的科技議題,主要訴求讓機器可以自動溝通、協調和設定,而不需「人」的介入,所以對於不同的系統或網路的穩定性要求都很高。因應這樣的技術,有愈來愈多的人力投入開發不同的智慧型服務應用,如國土防衛、消防安全,以及海洋、森林、農業等生態保護與產能開發,或一般家庭的老人看護、保全系統、節能綠屋等等,都屬於 M2M 的範疇。對開發人員來說,最重要的便是這些智慧型服務應用,必須提供什麼樣的服務,並有效的評估最適當的程序。AI 不僅可以用來處理底層的通訊與傳遞資訊,像是動態路由、自主連結,讓這些龐大的感測網路可以更有效地運用有限資源,AI 還可以設計來處理使用者的回饋,讓機器能理解使用者的需求,進一步地讓使用者的生活更加豐富且便利。 然而,要達到這麼多複雜的目的,最迫切且重要的議題是 ``如何讓系統可以根據使用者的需求,來重新調整機器的運作方式",如此一來,可以同時降低計算資源與人力工作的消耗,我們不需要再指派人力去重新設定這些 M2M 應用,而單靠機器彼此溝通協調就能達成。本論文提出了情境式系統配置代理人的 framework,來設計一個更具包容性及擴充性感測網路平台,降低佈建裝置和設定的人力工作。此 framework 包含三種提供給使用者們的文件: - Profile 用以描述機器的規格,讓安裝人員可以更方便地管理機器運作。 - Objective 用以描述系統的功能,讓開發人員可以擴充智慧型服務應用。 - Scenario 用以獲取一般使用者的需求及偏好。 因此我們只要用一套系統便能即時支援多樣的智慧型服務應用,且此系統具有擴充性,能容納大量的機器與設備同時運作。此 framework 還包含五種代理人,用來分析環境狀態和這些文件的資訊,並自動重新設定系統及電子設備的運作 (Reconfiguration) 和對環境自我調整 (Adaptation): - Network、Data、Transition 代理人分別負責網路訊號的溝通、資料的轉換與儲存、情境的自動判別。 - Reconfiguration 則處理情境的改變,要如何對應到機器的設定與操作。 - Adaptation 則選定適合回應的家電,提供使用者最好的服務品質。 本論文提出使用基因演算法來作 Adaptation,基因演算法可以幫助系統在執行過程中,自動因應情境變化來調整電子設備的運作方式,也可以讓新的設備進入時立刻和其他設備整合在一起。本論文最後以家庭和小型研討室的環境為例,來展示各種 M2M 應用以及使用者情境變化下,我們所提出來的 framework 可以因應隨時變動的環境和使用者的需求來調整。

並列摘要


In recent years, Machine-to-Machine (M2M) is the newest science and technology issue. The main purpose of M2M is to make machines could automatically communicate, coordinate and configure without human intervention. There are more and more intelligent applications were developed to serve people, such like elderly care and green houses. However, the most important issue is ``How to make the system reconfigure machines according to user needs?" As a result, we can reduce the waste of computing resources and manpower consumption. This thesis presents a scenario-based framework for a more inclusive and scalable sensor network platform. This framwork consists of three descriptive documents for users: - Profile is used to describe the specifications of the devices, so the installer can more easily manage the operation of machinery. - Objective is used to describe the functionalities of system, so the developer can expand the intelligent applications to serve more people. - Scenario is used to acquire the end-user needs and preferences. This framework consists of five agents to analyze the environmental contexts and these documents, then automatically reconfigure the system and devices operation, and adapt to the environment. - Network, Data, and Transition agents are responsible for the network communication, data collection, and the scenario identification. - Reconfiguration agent deal with the scenario changing and how to correspond to the setting and operation of the devices. - Adaptation agent select the approximate house appliances to provide users with the best services. The thesis presents the genetic algorithm (GA) for adaptation. In execution time, GA not only could help the system automatically in response to scenario changing and adjust the devices operations, but also immediately integrate the new devices with others. Finally, we use examples of home and a small seminar room, to show our framework is useful to adjust M2M applications according to the environment changing and user needs at anytime.

並列關鍵字

M2M adaptation reconfiguration

參考文獻


[1] J. M. Aaslyng, N. Ehler, and L. Jakobsen. Climate control software integration with a greenhouse environmental control computer. Elsevier Environmental Modelling and
Software, 20(5):521-527, 2005.
[3] P. Alexandros, M. Vangelis, S.-T. John, and M. Fillia. Sensor placement and coordination via distributed multi-agent cooperative control. In Proceedings of the 3rd
[4] T. Arampatzis, J. Lygeros, and S. Manesis. A survey of applications of wireless sensors and wireless sensor networks. In Proceedings of the 2005 IEEE International
sensor web. IEEE Pervasive Computing, 6(2):30-40, April 2007.

延伸閱讀