Sounds Good to Me

In the early days of the COVID-19 pandemic a very old technology was brought back into the spotlight — the face mask. The uncertainty about the severity and transmissibility of the disease led to widespread adoption of face masks among the general population for a time, where they had previously been a rare sight outside of healthcare facilities. All of this attention resulted in the development of many new mask-related innovations as people began to wonder how masks could be enhanced by incorporating modern technologies into their designs. These developments incorporated both sensors and processing units to do things like detect active infections, prevent risky behavior, or ensure a tight seal had been made.

For transparent integration of devices with a face mask, advancements in sensing technology needed to be made. Traditional, rigid chips that populate a circuit board are not always a good fit for an on-fabric device that comes into contact with the skin. Researchers at the City University of Hong Kong have added another tool into the mix for smart mask developers with their razor-thin nanocomposite sponge structure-based sound wave sensor. This flexible device is capable of detecting sounds like breathing, coughing, and speaking without being cumbersome or uncomfortable when installed in a face mask.

The spongy sound wave sensor is made of carbon nanotube/polydimethylsiloxane nanocomposites and is only 400 micrometers thick. It is highly sensitive in both static and dynamic pressure measurement ranges, and can also sense air movement, including directional flow and vibration. These characteristics make it useful for recognizing various types of respiratory activities. The lightweight, flexible sensor was integrated into a face mask by the team, and because of its design, it is virtually transparent to the wearer. Initially, machine learning algorithms were leveraged to detect different respiratory activities, but the team plans to expand this approach in the future to also diagnose respiratory diseases.

A small study was conducted with a total of 31 participants who wore a face mask outfitted with the researchers' smart mask. Each participant was observed while breathing, coughing, and speaking, and the mask was assessed for its ability to detect these activities. On average, the system was over 95% accurate in classifying each type of respiratory sound. These results are on par with what might be expected when using existing sound sensors, proving the utility of the researchers' device.

In addition to diagnosing respiratory ailments, the team also has other plans for their smart mask technology in the future. It has been noted that standard surgical masks influence the characteristics of one’s voice, which impairs communication with other people, and also with voice-controlled smart devices. By using their acoustic sensor and machine learning, they would like to develop a voice recognition algorithm that can offer assistance in both of these scenarios. It will be interesting to see what other types of applications will be developed with this novel sensor in the years to come, and also the types of platforms, aside from masks, it will be deployed on.