Researchers from the Institute of Microelectronics of the Chinese Academy of Sciences have developed a cutting-edge humidity sensing system capable of monitoring human behaviors in real-time through the detection of respiratory patterns. This innovative system combines a highly sensitive humidity sensor with a thermistor, micro-heater, and machine learning algorithms, achieving an impressive 96.2% accuracy in identifying various human behaviors. The technology holds large potential for transforming healthcare, smart homes, and daily health monitoring.
The system addresses the growing need for accurate and reliable human behavior recognition across various domains. Existing methods, such as video analysis and wearable devices, often present privacy concerns and environmental limitations. Respiration, a key physiological signal, varies with different physical conditions, making it a promising metric for behavior recognition. However, current humidity sensors lack the necessary sensitivity and stability to detect subtle respiratory shifts.
Published in Microsystems & Nanoengineering on January 22, 2025, the novel humidity sensing system employs porous nanoforests (NFs) as the sensing material. The sensor operates within a 60–90% relative humidity (RH) range and exhibits a sensitivity of 0.56 pF/%RH. A micro-heater enhances sensitivity by 5.8 times, enabling the detection of faint humidity changes in exhaled air. The thermistor ensures precise temperature monitoring, contributing to long-term stability and accuracy. With a rapid response time of 2.2 seconds and excellent gas selectivity, the sensor is well-suited for monitoring respiratory activity.
A convolutional neural network (CNN) analyzes the sensor’s humidity, temperature, and time data to recognize behavior. By converting one-dimensional signals into three-dimensional maps, the system accurately classifies nine common behaviors, including walking, sleeping, and exercising. The sensor, integrated into a mask, continuously collects respiratory data, wirelessly transmitting it to smartphones or computers for analysis.
Dr. Haiyang Mao, the lead researcher, emphasizes the system’s significance as a major advancement in real-time behavior recognition. The system’s combination of advanced sensor technology and machine learning creates a reliable and accurate tool for monitoring human behaviors, with profound implications for healthcare and smart home technologies.
The potential applications of this intelligent humidity sensing system are extensive. In healthcare, it can monitor patients with respiratory conditions or track physical activity levels. In smart homes, it can enhance comfort and safety by automatically adjusting appliances based on occupants’ behaviors. Furthermore, the system’s ability to detect subtle changes in respiration may provide valuable insights into emotional states, opening new pathways for mental health monitoring. With its accuracy and real-time capabilities, this system is poised to become a significant advance in health electronics and intelligent living.
The full research report can be found here.
Image: Microsystems & Nanoengineering
You must be logged in to post a comment.