智慧代理者基礎的PM2.5主動控制系統

Abstract

因為近年中國快速發展產生霾害問題，不僅讓台灣也開始重視霧霾的危害，如何控制空氣品質也成為重要的議題。台灣西部工業區與火力發電廠產生的懸浮微粒，因為地形與風向的關係擴散不易，工業區與火力發電廠周遭的居民也有罹患心血管及呼吸道疾病的風險。 本研究使用智慧代理者理論，透過空氣品質代理者收集本地的室內空氣品質資料，與環保署雲端伺服器提供的戶外空氣品質進行比對分析，並做出決策，啟動改善空氣品質的機制。其牽涉到雲端運算及服務、智慧代理者以及物聯網等3個範疇。 根據探討如何收集雲端資料庫資料與本地環境資料、雲端伺服器根據收集到的資訊做出決策以及改善空氣品質機制的方法的方法，本研究使用開放式機電系統Arduino來建構室內感測網路以及環境改善機構，圖控程式LabVIEW來接收資料、進行判斷動作以及啟動改善裝置，最後架設模擬環境並觀察本系統在模擬環境對於懸浮微粒濃度的改善情形來實做驗證。 本研究經實驗發現此系統在單一決策條件的環境下作用良好，但是對更多空氣品質條件的判斷時則會有衝突，需要撰寫更多的判斷式，或是設定動作優先權。未來可以將雲端決策系統架設網站並撰寫人機介面，用手機之類隨身上網設備就能隨時查看目前室內空氣品質以及控制狀態，也能增加感測器類別或是結合防災功能，利用更多條件改善環境品質。

Because the fast speed of China development in recent years causes haze problem, it not only let Taiwan begin to pay attention of haze hazards, controlling air quality also becomes an important issue. Industrial region and power plant in western Taiwan generate aerosols, what is hard to diffusion because of difficult terrain and wind direction, and the residents around industrial region and power plant at risk of developing cardiovascular and respiratory disease.

This study proposes the use of a control system of indoor air quality base on cloud decision to help user to control the concentration of suspended particles, it is a system of decision by comparing the data collected from local indoor air quality and the data of outdoor air quality provided by Environmental Protection Administration on the cloud server. The system involves the three category of cloud computing, wisdom and sensing network construction. After making a decision on the cloud server, the system can sent the command back to the indoor to switch on the smart building devices, to improve the indoor air quality.

By researching the method of collecting the data from local environment and the cloud sever, the method of decision making to improving the indoor air quality by cloud server, this study use the open-source electromechanical Arduino to build the sensor network and the environment improvement mechanism, and use the graphical programming software Labview to receive data, make decision and switch on the improvement devices. Finally, the study builds the simulation environment to observe the improvement situation of the concentration of suspended particles in it.

The study found that this system works well in the condition of a single decision-making role, but it may conflict when there are many air quality conditions to decide, and it needs more decision making method, or it needs to set priority action. In the future, there can be a website to upload the system with an user interface to monitor the indoor air quality and control it by phone. It can also add more sensor or combine disaster prevention, to improve the environment by more condition.