M2M通訊網路上即時監控之研究

Abstract

次世代M2M通訊技術儼然成為物聯網（Internet of Things）的發展基礎。無須人力介入，龐大數量的機器能夠互相連結用以創造具高度自主性及智能的應用服務。在眾多的M2M應用服務中，最具發展潛力的應用即為「即時監控」（Real-time Monitoring），其目的是透過區域式的M2M通訊網路來確保即時資料能被持續監控。在即時監控應用中，最重要的研究議題為傳輸耗能和即時資料正確性的權衡。為了解決這個議題，本論文提出省電傳輸機制來減緩機器傳輸資料的耗能，同時之間確保即時資料的正確性。本論文總共提出兩個傳輸機制，包含分散式傳輸機制（Energy-efficient Distributed Reporting; EDR）和集中式傳輸機制（Energy-efficient Centralized Reporting; ECR）。此外，本論文也提出分析模型從省電性和資料正確性兩方面來研究傳輸機制之效能。再者，在本論文中，我們實作空氣汙染感測系統（Air Pollution Sensing System; APS）。APS系統為即時監控系統之雛形，用以監控都市區域的空氣品質。此系統已經建置於台灣大學中進行小規模的實測。

Machine-to-Machine (M2M) communications have emerged as a new communication paradigm to support Internet of Things (IoT) applications. With little or even no human intervention, the interconnection among a large amount of machines enables highly autonomous and intelligent applications. Undoubtedly, the most promising M2M application is ``real-time monitoring', the goal of which is to ensure continuous surveillance via region-based M2M communications networks. One major issue for real-time monitoring is the energy-validity tradeoff, i.e., the tradeoff between the energy efficiency for machines and the validity of sensing reports. To deal with the energy-validity tradeoff for real-time monitoring applications, in this thesis, we propose energy-efficient reporting mechanisms, including Energy-efficient Distributed Reporting (EDR) and Energy-efficient Centralized Reporting (ECR), to mitigate energy consumption for machines while keeping the sensing data valid. We propose analytical models and simulation experiments to investigate the performance for the proposed mechanisms in terms of the power saving and the report validity. Furthermore, we develop a prototyping real-time monitoring system, namely the- Air Pollution Sensing (APS) system, to monitor the air quality in the urban environment. The APS system has been deployed in the campus of National Taiwan University for small-scale testing.