情境感知之綠色智慧生活平台

Abstract

為了追求更健康、安全與舒適的生活同時降低資源的消耗，智慧家庭的概念已經被提出幾十年了。智慧家庭係指整合了各式服務的生活環境。雖然各種的科技已經被發展出來，但是有關智慧生的科技卻沒有普及。 造成這樣的情形有三個因素。第一，通常安裝與設置一個智慧家庭的系統是極為複雜的。一般使用者無法輕易的自行安裝與設定。第二，時不時的要去維護系統內的裝置對使用者造成及大困擾。任何維修或者重新繞線都會造成時間上的消耗與金錢上的負擔。第三，在不侵犯使用者隱私的前提下用來提供情境感知服務的資訊難以取得。一旦使用者覺得隱私被侵犯他將不會繼續使用該服務或系統。 根據以上觀察，一個新的平台在這篇論文中提出。直覺是的操作與設定機制讓使用者可以輕鬆的完成系統的安裝與設定。在系統中的裝置還能自行的組成控制連結，完全不需要使用者做手動的設定。一系列的非侵入式感測裝置也在這篇論文裡被提出來，用以偵測使用者行為。這些感測裝置由於不需要任何外部電源來進行運作，因此他們一旦安裝後便不需要進行維護。除此之外，開發者更可以在這個平台上利用所提供的 API 開發新的服務。因為異質的裝置得以透過這個平台相互通訊，跨範圍整合服務得以被提供。 為了展示這個平台的可行性，這個平台被實際安裝於一間位於新北市的公寓當中。一系列的服務，包含：日常生活監控、安全警報、家電使用監控與智能燈光控制等服務被架設於這個平台之上。結果顯示這個新的平台式非常容易安裝設定、不需要太多的維護而且可以提供許多跨範圍的服務。

To pursue a healthier, safer and more comfortable life while reducing resource consumption, the concept of smart livings or smart homes have been proposed for decades. Smart homes are living environments that integrate various services to enhance the living quality of the residents. Even though various technologies have been developed, the mass adoption of smart-living technologies and services has been very slow. Three key factors have led to such slow adoption. First, the installation of a smart home system is usually complicated. It is not easy for ordinary users to configure the system in their homes by themselves. Second, maintenance is always a headache. Any repair that requires rewiring or changing batteries is not only time consuming but also very costly. Finally, the information needed to provide context-aware services is hard to obtain without violating user privacy. People prefer not to use the services if they think their privacies are invaded. Based on the above observations, a novel platform for smart living is proposed in this thesis. Intuitive control and automatic configuration mechanisms are provided for users without any technical background to set up and configure the system. Furthermore, the devices in the platform are able to automatically establish control links between each other and adjust their settings. A full range of non-intrusive sensing devices including self-powered sensors and radio sniffers are developed to collect detailed user behavior information without directly invading user privacies. Since the self-powered sensors do not need any outer power source to function, they are easy to install and do not need maintenance. Other than the devices originally supported in the platform, new devices and services are able to be deployed using the APIs provided by the platform. Because heterogeneous devices are interconnected, the platform is able to perform cross domain device cooperation and big data collection to provide customized integrated services. In order to demonstrate the feasibility of the proposed platform, a smart living space based on the proposed platform is created in an off-campus, two-bedroom apartment in New Taipei City. Several services including daily activity monitoring, safety alert, appliance usage monitoring, and autonomous lighting control are developed based on the platform. The results show that the proposed platform is 1) simple to install and configure, 2) dramatically reduces the maintenance process, and 3) is able to provide cross domain, integrated services without making users feel that their privacies are directly invaded.