A Real-World Smart Home Dataset Integrating Sleep, Environmental, Physiological, and Ambient Sensing for Homecare Research

Abstract

Modern homecare research increasingly relies on multimodal sensing technologies in smart homes to monitor daily routines, sleep, environmental conditions, and physiological activities over long periods. However, publicly available datasets often lack real-world longitudinal tracking, multimodal integration, and detailed environmental and wearable sensor data collected in real home environments. This paper introduces a multimodal smart home dataset, collected from 10 participants over approximately one month in a real-world residential environment. The dataset integrates various Internet-of-Things (IoT) sensing modalities, including motion sensors, door contact sensors, environmental sensors, wearable physiological monitoring devices, and under-mattress sleep tracking mats. The collected data includes timestamps of room occupancy, steps, sleep, heart rate, respiratory rate, snoring, temperature, humidity, and interactions within the home. All sensor data streams were represented using timestamps standardized to UTC and stored as structured event records, including user and sensor ids, measurement types, timestamps, and sensor values. The dataset was then organized into CSV (comma-separated values) files for each user to facilitate future research in areas such as smart home data analytics, behavior monitoring, sleep analysis, home assistance, homecare systems, activity recognition, anomaly detection, and digital twin-based homecare applications. This dataset aims to support the reproducibility of research in homecare and ambient assisted living (AAL) environments.