Methods of human activity classification in buildings

Abstract

The number of smart homes is rapidly increasing. Smart homes typically feature functions such as voice-activated functions, automation, monitoring, and tracking events. Besides comfort and convenience, the integration of smart home functionality with data processing methods can provide valuable information about the well-being of the smart home residence. This study is aimed at taking the data analysis within smart homes beyond occupancy monitoring and detection of falling events of people. Two different approaches are proposed to integrate human activity recognition within smart homes. The first approach utilizes KNX standard-based devices to obtain room air quality data (humidity, CO2, temperature) and combine the obtained data with two wearable devices that provide movement-related data. The second approach simplifies, improves, and addresses a few of the shortcomings of the first approach, it utilizes different measuring devices with higher sampling rates. It examines multiple statistical methods and ultimately chooses a simpler multi-layer perceptron neural network model. Resulting in a less computationally intensive solution with higher accuracy levels. The study achieved cross-validation accuracy levels above 98 %.