基於非侵入式負載監測電錶之電器狀態與使用者行為對應

Abstract

在這篇論文中，我們設計了一個能源管理系統，其中包含了非侵入式負載監測(nonintrusive load monitoring)電錶，閘道器，伺服器與使用者界面裝置。非侵入式負載量測電錶可辨識出電器的種類與狀態及量測其耗電資訊，以OSGi平台為基礎的閘道器為伺服器與建立在Android平板上的使用者界面提供服務。Android應用程式包含了四個功能:設定電器的位置與資訊，展示電器的及時狀態與以圖表的方式呈現耗電的歷史資訊，住戶可設定每個月的最大耗電量以管理電器的使用，和及時或設定某個時間對電器執行開關的動作。

除此之外，我們在能源管理系統之上加入一個主動分享個人行為的應用服務。傳統使用者行為的偵測大多仰賴大量的感測器與複雜的監控設備，價格昂貴、部建不易、且侵入式的方式讓人較難接受。在此篇論文中，我們考慮以使用者用電(或是使用電器)的足跡加以偵測使用者行為，並透過非侵入式負載監測電錶偵測電器狀態改變，我們利用電器狀態改變與使用者標記的行為時間點找出它們之間的序列模式(sequential patterns)，並用此序列模式來偵測出使用者的行為。

In this paper, we design and implement an energy management system (EMS) consisting of a nonintrusive load monitoring (NILM) smarter meter, gateway, server and mobile device. The NILM smarter meter provides a low-cost solution to recognize the states of each appliance in the house and disaggregate the power consumptions of a house. The gateway is based on Open Service Gateway Initiative (OSGi) framework to provide services for servers and mobile devices such as Android pad. The Android application consists of four components: the settings set configuration of appliances and location, the monitor used to show the state of appliances in real-time and the charts of historic data, the management can set the maximum power consumption of a month by the inhabitants, and control used to control the appliances now or at the time that inhabitants configured. Based on the proposed EMS, we further implement a user behavior sharing service. Conventional user behavior detection relies on a large amount of sensors, complicated and expensive monitoring devices. However, such monitoring system suffers from deployment problem and intrusive monitoring concerns. Therefore, we consider detecting user behavior based on NILM smart meter which provides us a sequence of appliance usage patterns. We process sequential patterns of appliance usage and derive high level users’ behaviors.