Development of a Miniature Smart Home Testbed for Research and Education.
Abstract
We present a cloud based smart home testbed using a miniature doll house. In particular, we use this testbed to demonstrate the energy conservation system with the interconnected network of devices, the robust security system with face recognition feature, and the home assistant robot with voice-controller. The test bed consists of a smart home server, a home controller, a smart home assistant robot, a security camera, and appliances communicating with each other using web-socket and existing Social Network Software SDKs. The proposed testbed allows research and education in areas such as smart-grid, wireless sensor networks, machine learning, pattern recognition, embedded programming, natural language processing, social media sharing etc. We also propose a security system based on face recognition. In particular, we develop this system for giving access into a home for authenticated people. The classifier is trained using a new adaptive learning method. The training data is initially collected from social networks and the accuracy of the classifier is incrementally improved as the user starts using this system. A novel method has been introduced to improve the classifier model by human interaction and social media. By using a deep learning framework - TensorFlow, it will be easy to reuse the framework to adapt with many devices and applications.
Collections
- OSU Theses [15752]