多穿戴式裝置於短距無線傳輸之輪詢設計

A General Polling Design for Multiple Wearable Devices Using Short-range Wireless Communications

陳瀅婷(Ying-Ting Chen)

指導教授 ： 蔡志宏

國立臺灣大學/電機資訊學院/電信工程學研究所/碩士(2015年)

https://doi.org/10.6342/NTU.2015.10589

摘要

近幾年來，由於人們對個人健康照護與運動的重視，越來越多的穿戴式裝置出現在市場上，許多相應的應用也為改變人們的生活而生。然而，當這麼多穿戴式裝置同時被運用時，一個重要的技術課題是如何讓這些裝置即使遭遇不同應用而有不同的記憶體限制、封包與資料傳輸速率，而仍然在現有的短距無線通訊技術下確保所有裝置的封包都能在延遲上限內抵達。

本論文中，我們討論以智慧手持裝置作為所有穿戴式配備連網核心的網路架構。我們希望可以藉由針對穿戴式身軀區域網路下封包週期性抵達的特質，修改既有的輪詢系統分析與設計輪詢的序列來解決相關的服務品質(QoS)問題。我們提出了一個相應於此網路的系統模型並謹慎的進行系統分析，並探討實作此系統的四個主要部分: 通訊型態通訊裝置的硬體實作、藍芽通訊協定在硬體的實現方式、通訊型態通訊裝置的作業系統以及行動裝置上的應用程式實作方法。以上這些觀念、分析、方法和設定在本論文中均予以詳細討論。另外，我們也會呈現每個部分的測試結果。

本論文提出一個名為通用輪詢設計(GPSD)的演算法並簡易的在Android裝置上實作了一個基礎的輪詢模型。通用輪詢設計包含三個主要的部分: 系統檢查、找到所有優質的序列並找出符合需求的候選者以及序列選擇。系統檢查針對裝置的組合檢查了所有的穩定度條件以及其他基本的系統要求。完成這些檢查項目後，手持裝置可以開始尋找所有可能優質序列並確認其中可以符合所有裝置延遲限制之序列。手持裝置再選擇一個對最高優先保護的裝置是有最高延遲保障的序列來確保最需要被保護的裝置不會遺失封包。最後本論文經由模擬分析，驗證同一演算法，在不同情境下與不同演算法的一般設定做比較。由分析結果得知，本論文提出的演算法可以有效的改善使用者體驗並讓使用者可同時的使用更多裝置。並且，此演算法的實作是簡易且可行的。

關鍵字

穿戴式 ； 多裝置 ； 輪詢 ； 身軀區域網路

並列摘要

As demands for personal healthcare and workout training grow, a huge number
of wearable devices have come to the market in recent years. Many devices and
applications have been designed and developed to change the way people live.
Services which didn’t exist in the past seem to be available and are discussed lively.
One of the important issue is how to guarantee the delay requirements of devices
when they are having different buffer limits, packet size, and data rates, using short
range wireless communications.
In this thesis, we focus on the network architecture with a smart handheld as
center of wearable devices collecting sensing data in a wireless wearable body area
network. We aim to solve related QoS problems by refining the polling model under
periodic processes and designing a polling sequence for given devices. We propose a
system model for the network and made careful analyses on this architecture,
including hardware architecture on MTC devices, introduction of Bluetooth protocol
stack, operating system on MTC devices and implementation methods on mobile
central devices. The concept, analyses, methods and setup in this thesis will be
described, and testing results of each component are demonstrated.
II
We proposed an algorithm called General Polling Sequence Design (GPSD) and
implemented a related basic model of it on Android devices. GPSD contains three
main parts: System Checking, Finding of Possible Sequences and Candidates, and
Sequence Selecting. The System Checking checks the stability condition and other
basic system limit for given devices. After this checking, the handheld can start the
Finding of Possible Sequences and Candidates and know that whether there is a
sequence satisfies all delay requirements for devices. The handheld can then select
one sequence with highest delay tolerance for high-priority device among them to
make sure its packet will not be dropped.
The proposed algorithm, GPSD, was validated and evaluated under several
scenarios via simulations. We also compared the results with the default systems. By
simulation results, GPDS could improve user experiences and accommodates more
devices and services. Moreover, the implementation of it can be easy.

並列關鍵字

wearable device ； multiple device ； polling ； body area network

參考文獻

[1] N. Johansson, U. K ouml;rner, and P. Johansson, "Performance evaluation of scheduling algorithms for Bluetooth," Broadband communications, Springer US, 2000.

[2] R. Palacios, et al, "Analysis of an Energy-Efficient MAC Protocol Based on Polling for IEEE 802.11 WLAN," IEEE International Conference on Communications (ICC), 2015.

[3] A. Abdrabou, and A. M. Gaouda, "Considerations for packet delivery reliability over polling-based wireless networks in smart grids," Computers Electrical Engineering, 2014.

[4] Y. Li et al, "Energy modeling and optimization through joint packet size analysis of BSN and WiFi networks," IEEE 30th International Performance Computing and Communications Conference (IPCCC), 2011.

[5] A. Awad, et al, "Real-time implementation and evaluation of an adaptive energy-aware data compression for wireless EEG monitoring systems," IEEE 10th International Conference on Heterogeneous Networking for Quality, Reliability, Security and Robustness (QShine), 2014.