適用於動態無線近身網路媒體存取控制層的通道模型

Abstract

無線近身網路(WBAN)近來年在醫療照顧以及醫學應用上受到重視。在這深具潛力的無線網路系統領域中，目前已經有許多通道模型被提出。然而，這些被提出的通道模型大部分都是以物理層的角度來提出的，且當中鮮少是以動態的量測為基礎。 在本論文中，我們提出了一個專為媒體控制層所設計的動態通道模型。我們量測兩個重要的行為，步行與睡眠。除了考量所接收訊號振幅的統計特性之外，我們另外也研究這二種行為各自在時間上的相關性來強化在媒體控制層的模型準確度。我們提出了雙階層模型以及三層次模型來分別特性化這兩種行為。藉由在模型中考量時間的相關性，所提出的模型可以在媒體存取層達到相當高的模型準確性。

Wireless body area network (WBAN) has been paid attention in health-care and medical application field in recent years. Many of the channel models have been proposed for this potential wireless network system. However, most of them are built in physical layer, and they seldom analyze the body channel in dynamic scenario. In this thesis, we propose a MAC channel model for dynamic WBAN. We perform the dynamic measurement of two vital behaviors, walking and sleeping. Apart from considering the conventional statistics of received power amplitude, the model also investigates the time-domain correlation of those two activities to enhance the modeling accuracy in MAC. We propose the two-state model and three-level model to characterize the two behaviors, respectively. With the consideration of time-domain correlation in modeling, the proposed models can achieve high accuracy with the measured data in MAC point of view.