多人無線近身網路系統

Abstract

無線近身網路可利用短距離之無線傳輸擷取並傳送人體之重要生理資訊。其目前已被視為實現遠距照護(無遠弗屆醫療照護)之最重要的”最後一公尺”。無線近身網路其架構可以相當多元。透過適當的假設，無線近身網路可被比喻為”個人化之感測網路(personal sensor network)”、”社群用之隨意網路(social ad-hoc network)”、或是”醫療用之網狀網路(medical-specific mesh network)”。也因此目前許多既有之感測、隨意、與網狀網路之技術可用來增進無線近身網路之效能。本研究進一步探討無線近身網路特有之”近身網路間干擾”議題。此議題來自於無線近身網路使用者之移動性。經常性的近身網路移動所導致之彼此重疊或碰撞造成無線節點密度之劇烈變化，因此引發劇烈之近身網路間互相干擾而導致傳輸效能與可靠性之低落。透過建立多近身網路使用者之干擾模型，本研究率先指出近身無線網路之設計挑戰。其包含: 1) 在高密度近身網路情境下，如何同時增進網路使用效能與快速反應因使用者移動性所帶來之拓樸與干擾變化?(以傳統圖學資源分配觀點來看，此兩校能參數互為取捨) 2) 如何在多使用者近身網路中建立洽當之優先權制度?此制度須同時考量感測器類別、使用者狀態、與緊急事件總類等之個別差異性。 3)如何在分散式網路的架構下實現以上需求? 在本論文中，我們針對各議題做深入討論並提供其個別的解決方案，包含隨機非完整著色法(Random Incomplete Coloring)用來最大化無線近身網路資源分配之速度與頻道利用效率與隨機資源競爭(RACOON)演算法用來實作分散式近身網路之優先權制度。其個別效能探討將搭配相關理論模型與電腦實驗。

Wireless Body Area Network (WBAN) is designed to retrieve important body signals through short range wireless technology in charge of the “last meter” of ubiquitous healthcare. With proper assumptions of network architecture, WBAN can be regarded as a “personal sensor network”, a “social ad-hoc network”, or a “medical-specific mesh network”. Based on the experiences of these well-known wireless technologies (sensor, ad-hoc, and mesh networks), this study further discusses and puts more focus on effects of “inter-WBAN interference.” This multi-WBAN-specific issue is caused by the mobility of WBAN users. When WBANs are moved with users, they frequently encounter and overlap with each other. These WBAN-overlaps lead to dramatic changes of sensor density and hence creates dynamic mutual interferences between WBANs. Through establishing interference models of multiuser WBANs, this study first identifies design challenges under the inter-WBAN interference. Primary challenges include 1) simultaneously improving channel efficiency in high density WBAN while maintaining fast response time to rapid WBAN topology changes (these two performance indexes are tradeoffs in term of traditional scheduling theory), 2) Constructing the proper priority scheme for multiuser WBAN which can simultaneously satisfy priority requirements from sensor types, user conditions, and emergency-handling points of views, and 3) solving these problems in distributed WBAN systems. This study will discuss above issues in separate chapters. Corresponding solutions, Random Incomplete Coloring (RIC) and RACOON algorithms that solve speed-efficiency trade-offs of traditional coloring and WBAN-specific priority schemes, respectively, will be introduced. Improvement of related performance indexes will be carefully verified by proper analytical models and experiments.