FaceBit, the “Fitbit for the Face,” Attaches Smart Sensors to Masks

January 19, 2022 by Darshil Patel

The umbrella of wearables just broadened even wider with face masks. Engineers from Northwestern University have developed an embedded hardware platform that can sense users’ biometrics through a face mask.

Since face masks are such a critical tool to combat COVID-19, researchers are using emerging technologies to make them more effective. A team of engineers from Northwestern University (NWU), led by Dr. Josiah Hester, have developed a new sensor-based hardware platform for face masks. This "Fitbit for the Face" or "FaceBit" is a lightweight, quarter-sized sensing platform that can turn any facemask into a smart monitoring device. The hardware uses a tiny magnet to attach to any mask.


A FaceBit device placed in an N95 mask via a magnet

A FaceBit device placed in an N95 mask via a magnet. 


The researchers recognize that face masks are well-positioned to give valuable health metrics. With the active protection of smart face masks, users can get notified if the mask is not worn correctly or whether there's a leak. It can also function as a wearable health tracker, its position over the mouth giving it unique advantages over other wearables.


What Can a Smart Mask Measure?

The developed hardware platform can sense users' real-time respiration rate, heart rate, mask wear time, and misfitting of the mask. All this data is transmitted to a smartphone app that includes an analytics dashboard. The app alerts users when issues are detected, such as elevated heart rate or a leak in the mask.


The working principle behind FaceBit

The working principle behind FaceBit. 

Realizing the hardware for such smart face masks was a challenge. Smart active protection masks have been proposed previously, but they are expensive, weighty, and high maintenance. Because masks must constantly be replaced—either to dispose or wash—they must be inexpensive and easy to wear. The NWU researchers recognized a market gap for a system that could transfer existing data from one mask to another. 

They addressed these concerns by developing low-cost and energy-efficient hardware. This sensor-based system can be used in any mask without modification and can easily be clipped onto any face mask using a magnet.


FaceBit Power Architecture

Wearable systems require a battery large enough to last a week or longer. But this necessity adds to the weight of the system and may cause discomfort to the user.

FaceBit employs a hybrid approach when it comes to power. It includes a battery holder for a small 105 mWh cell along with energy harvesting circuitry and storage devices for powering the board via both DC and AC power generation sources.

It integrates three tantalum capacitors of 3mF that can hold a total of 104 µWh of harvested energy. These capacitors are charged by Texas Instrument's BQ25570 power management IC, which features a boost converter with MPPT, a charging circuit, and a buck converter for voltage regulation. 


Flexible hybrid power architecture of FaceBit

Flexible hybrid power architecture of FaceBit. 


The hardware is primarily powered by the battery. When the capacitors' voltage exceeds 3 V, the battery boost converter is disabled and the system draws power from the capacitors. When the capacitors' voltage falls below 2.6 V, the system switches to draw power from the battery.

According to the researchers, this architecture allows for longer battery life while maintaining the ability to store and use harvested energy. The device also includes additional flexible modes, such as a mode that makes the system only run from capacitors' stored energy.


Platform System Design and Implementation

FaceBit features Nordic Semiconductor's nRF52832 Bluetooth 5.3 SoC within u-blox's BMD-350 Bluetooth low energy (BLE) module. The current version of FaceBit uses six sensors. 


Top (left) and bottom (right) of the FaceBit board

Top (left) and bottom (right) of the FaceBit board.


FaceBit relies on sensors and firmware to gather data about health metrics. The firmware takes measurements when the mask is being worn; the board uses only a barometer to record samples, keeping an eye on the minimum and maximum pressure values. If the difference between these two values is greater than 0.1 mbar, the system determines that the mask is being worn.

Heart rate is measured using a ballistocardiograph (BCG). The recoil generated throughout the body with each heartbeat is detected by sensitive accelerometers, like the onboard LSM6DSC inertial module from STMicroelectronics. The measured pulse waves consist of several high-frequency waves, which are then processed by various digital filters and other signal processing tools.

The respiratory rate is measured by onboard pressure sensors as the pressure drops over the filter material during normal breathing. The researchers note, however, that loose-fitting masks do not generate such large signals.

All the data is transmitted wirelessly via the BLE module using a custom GATT profile and is displayed in the FaceBit companion app developed for iOS and macOS. The app consists of various sections for the user and for healthcare providers.


FaceBit companion application for iOS

FaceBit companion application for iOS. 


What's Next for FaceBit?

To understand the capability of FaceBit, the researchers conducted a small-scale study with the help of volunteers and improved the algorithms as problems arose. According to NWU, the board still needs to undergo clinical trials. 

In the future, the researchers are aiming for FaceBit to run only on harvested energy. In upcoming studies, they are planning to reduce false detections and make the system more robust. In addition, the researchers are aiming to reduce the number of sensors and thereby reduce the size and cost of the device.

The work is published in Proceedings of the ACM on Interactive, Mobile, Wearable, and Ubiquitous Technologies. The entire system is open-source, allowing creative designers to build on what NWU researchers have developed. The hardware design files and source code for firmware and the app are available on the project's GitHub repository.



All images used courtesy of Northwestern University