The method for hierarchical context awareness based on activity recognition

we propose a hierarchical context awareness method based on human activity recognition (HAR). More specifically, the smart phone collects sensor data including wearable devices and peripheral IoE sensors, classifies the user’s simple actions based on machine learning, and transmits only the changed context information to the server to provide quick context aware service. At this time, in order to provide a service suited to the situation, the smart phone can provide connection request and connection reconfiguration with peripheral devices. The server provides high-level context awareness based on transmitted context information and various knowledge base data. At this time, knowledge base data can be newly generated based on machine learning, and context information history can be used to extract user’s behavior pattern, Situation prediction, and personalized behavior analysis model.