With the development of Internet of things (IoT) technology and the increasing demand for building energy management, many buildings are adopting intelligent building management systems based on IoT devices. The intelligent building management system not only collect large amounts of real-time data from IoT devices, but also require data analysis to monitor building environment. This paper proposes a data processing framework for IoT-based intelligent building management (IBFRAME). The proposed framework aims to provide data collection, monitoring, processing, routing, and analytics services to support a variety of intelligent building applications. In this framework, an ingress router converts and forwards incoming data to application servers and data processors. To monitor building environment, data enrichers analyze thermal comfort and occupancy detection in each space of the building based on IoT sensor data, such as temperature, humidity, and CO2concentration. Data collectors of the framework gather different types of data from external data sources, such as weather databases and energy meter devices. This framework is used as the foundation for processing and analyzing IoT data in intelligent buildings, and designed to facilitate module extension. It supports the implementation and operation of diverse intelligent building application services. This paper also discusses the structure of the system that is built and operated based on the proposed framework. The system adopts a container-based time series database, which specializes in massive time series data processing, as a data management system.

Retrofit could be still expensive solution to existing building. However, with IoT(Internet of Things) environment sensors, it is possible to build a powerful intelligent building monitoring system with relatively low costs. The IoT environment sensor can analyze the IAQ(Indoor Air Quality), occupancy, and comfort level of the individual space by integrating temperature, humidity, carbon dioxide, sound, illuminance, organic compound, human body detection and IR sensor in the individual space of the building. In addition, because it supports Wifi mesh network, it can be easily installed without performing construction for retrofit in existing building, and real time monitoring can be performed. In this paper, we introduce the result of applied the IoT environment sensors to three buildings and visualization interface based on the result of analysis and real environment sensing

In order to utilize a unmanned aerial device such as a drone in terms of reconnaissance, surveillance and disaster management, a technique capable of performing real-time human detection is needed. In this study, we introduce a method of human detection by constructing environment similar to a drone by installing wide angle lens camera on the ceiling. Tiny-Yolo, which is a kind of Convolutional Neural Network, was used as recognition technology. Camera sensors were installed on ceilings in various places inside the building to collect training data for human detection. The Training was performed based on 4,000 pieces of collected data, and the accuracy of recognition was measured. As a result, 95.37% precision and 96.60% recall were obtained.