ÌìÃÀ´«Ã½

Newswise — Human behavior recognition has become increasingly vital across various domains, from healthcare to smart home automation. Traditional methods, such as video analysis and wearable devices, often face privacy concerns, environmental limitations, and the need for multiple sensors. Respiration, a key physiological signal, changes with different physical conditions, making it a promising metric for behavior recognition. However, current humidity sensors fall short in terms of sensitivity and stability, particularly when detecting subtle respiratory shifts, such as rapid or weak breathing. This gap has highlighted the urgent need for advanced sensors capable of accurately tracking and analyzing human behavior in real-time.

In an exciting development, researchers from the Institute of Microelectronics of the Chinese Academy of Sciences have introduced a novel humidity sensing system, in on January 22, 2025. This system incorporates a thermistor and micro-heater with porous nanoforests as the sensing material, achieving an impressive 96.2% accuracy in recognizing human behaviors through respiration monitoring. The integration of machine learning further enhances the system's ability to provide real-time analysis, setting the stage for transformative applications in healthcare and smart home technologies.

At the heart of this research is the innovative humidity sensor utilizing porous nanoforests (NFs). The sensor operates within a humidity range of 60–90% relative humidity (RH) and boasts a sensitivity of 0.56 pF/%RH. A micro-heater enhances its sensitivity by 5.8 times, enabling the detection of even the faintest humidity changes in exhaled air. The inclusion of a thermistor allows for precise temperature monitoring, ensuring long-term stability and accuracy. With a rapid response time of just 2.2 seconds, along with excellent gas selectivity, the sensor is ideally suited for monitoring respiratory activity.

Behavior recognition is driven by a convolutional neural network (CNN) that analyzes the sensor's humidity, temperature, and time data. By converting these one-dimensional signals into three-dimensional maps, the system can classify nine common behaviors, such as walking, sleeping, and exercising, with a high degree of accuracy (96.2%). Integrated into a mask, the sensor continuously collects respiratory data, which is wirelessly transmitted to smartphones or computers for analysis. This seamless fusion of hardware and software demonstrates the system's immense potential for practical use in healthcare and daily life.

Dr. Haiyang Mao, the lead researcher of the study, emphasized the significance of this breakthrough: "This innovative humidity sensing system represents a significant leap forward in real-time behavior recognition. By combining advanced sensor technology with machine learning, we've created a reliable and highly accurate tool for monitoring human behaviors, which has profound implications for both healthcare and smart home technologies."

The potential applications of this intelligent humidity sensing system are vast. In healthcare, it could be used to monitor patients with respiratory conditions or those needing to track physical activity levels. In smart homes, it could enhance comfort and safety by automatically adjusting appliances based on occupants' behaviors. Furthermore, the system's ability to detect subtle changes in respiration may also provide valuable insights into emotional states, such as anxiety or stress, opening new pathways for mental health monitoring. With its impressive accuracy and real-time capabilities, this system is set to be a cornerstone of future health electronics and intelligent living.

###

References

DOI

Original Source URL

Funding information

This work was supported by National Natural Science Foundation of China (Grant Nos. 62474192 and 62201567), Youth Innovation Promotion Association, Chinese Academy of Sciences (Grant Nos. 2022048 and 2022117), State Key Laboratory of Dynamic Test jointly built by Province and Ministry Open Fund (Grant No. 2022-SYSJJ-07).

About 

 is an online-only, open access international journal devoted to publishing original research results and reviews on all aspects of Micro and Nano Electro Mechanical Systems from fundamental to applied research. The journal is published by Springer Nature in partnership with the Aerospace Information Research Institute, Chinese Academy of Sciences, supported by the State Key Laboratory of Transducer Technology.