於示範中萃取事件-狀態-行動規則以編程物聯網之研究

Abstract

物聯網將生活中的物品連上網路，帶來了資訊分享與操控的便利性。然而，現有的編程方法仍然阻礙終端使用者利用這些好處。如何能讓使用者不用程式語言或是訓練學習模型，我提出示範物品之間的實際運作，以及背後探勘隱藏的模式來編程物聯網。本論文描述一個系統從感測器和控制器收集之資料萃取出事件-狀態-行動規則為主的程式，以及與狀態機計算模型的關聯。最後本論文透過比較實際案例來證實系統的實用性與適用範圍。

The Internet of Things (IoT) connects massive daily things to conventional Internet networks, bringing the benefit to get information from or control those things with ease. However, current programming solutions still hinder the end-users to take advantage of the emerging technology. Instead of asking them to write programs or to train models, I proposed to program IoT by demonstrating how the things should work explicitly and combine the frequent patterns discovered implicitly. This paper presents a system to generate Event-Condition-Action rules from the data collected from the sensors and the actuators and relate the mining to the computation model, Finite State Machine. The results and comparison study are presented to show the feasibility and the limitation of the system.