IoT-enabled smart mask for monitoring body parameters and location through cloud

Abstract

In the last interim guidance, the WHO advised the use of masks in communities, during home care, and in healthcare settings in areas with reported cases of COVID-19. This advice was intended for individuals in the community, public health and infection prevention and control (IPC) professionals, healthcare managers, healthcare workers (HCWs), and community health workers. As the two primary routes of transmission of the COVID-19 virus are respiratory droplets and contact, face masks have become potential safety tools in public places. Subsequent contact with the face, eyes, nose, and mouth following contamination is detrimental. However, during this pandemic, physicians and nurses have suffered the consequences of wearing face masks for several hours. Therefore. a full-face mark, which ensures filtered breathing during the day, is critical. The proposed smart mask model protects the respiration of mask wearers and monitors their body temperature, sneezing attacks, and social distancing. Moreover, it registers their geographical location after ensuring ID registration. The proposed mask can be used both indoors and outdoors. Moreover, data can be processed locally for alarms related to temperature and social distancing. The remaining data are sent to a cloud for post-processing to record the histories of mask wearers in all parameters, including their geographical dynamic location, to track the possible spread of contamination. The prototype and measurement results demonstrate the practicality and potential utility of mass numbers.

1. Introduction

On 6 April 2020, the WHO issued interim guidance, advising on the use of masks in communities, during home care, and in healthcare settings in areas where cases of COVID-19 have been reported [1]. This advice is intended for individuals in the community, public health and infection prevention and control (IPC) professionals, healthcare managers, healthcare workers (HCWs), and community health workers. Current investigation reveals that the two primary routes for the transmission of the COVID-19 virus are respiratory droplets and contact. Respiratory droplets are produced when an infected person coughs or sneezes. Anyone in close contact (within 1 m) with such a person is at risk of being exposed to infectious droplets. Droplets can also land on surfaces where the virus remains viable. Thus, the immediate environment of an infected person can serve as a source of transmission (contact transmission). Touching the face eyes, nose, and mouth after contacting such an environment is detrimental.

Owing to the long hours of wearing face marks during this pandemic, physicians and nurses have suffered damage to their faces. Therefore, a full-face mask, which enables filtered breathing during the day, is critical. Thus far, we believe the proposed smart mask is the best alternative solution.

To address the ongoing COVID-19 pandemic, the US Center for Disease Control and Prevention (CDC) has encouraged the wearing of face masks [2]. Several Asian countries where viral illnesses frequently occur use public masks. Several countries worldwide have announced policies mandating the wearing of covers to reduce the spread of COVID-19. Mandatory use and implementation vary internationally. Although some nations have enacted regulations mandating the wearing of masks, others, such as China, India, Japan, South Korea, and Taiwan, have enacted more specific regulations.

COVID-19 protective masks can be classified into conventional and smart masks. Conventional masks are typically used in hospitals, with some extensions. These include cloth masks and respirators. Smart masks can protect against the spread of the virus. Moreover, they can detect any COVID-19 virus infection or monitor the vital signs of the person wearing the mask.

Masks equipped with biosensors detect the presence of viral infection using biosensors woven into the mask [1,2]. A biosensor using a freeze-dried organic apparatus, a CRISPR1 -based hereditarily designed circuit, and fluorescent particles have been developed. Biosensors can provide a perceptible sign based on the presence of a target molecule. The eventual product comprises three distinct organic responses that are successively actuated when the pressing of a button discharges water.

The other type of smart mask monitors vital signs through various combinations of electronic sensors and sends data to a remote server for further processing. The smart mask designed in this study falls into this category.

In this type of application, smart materials with embedded sensors [3] can be used for online monitoring in the context of Industry 4.0 to collect and further process data automatically [4]. This is of interest for IoT applications in condition-based monitoring with all types of data collection [5]. Despite extensive studies, masks that include temperature sensors are required to measure human body temperature and sneezing, in addition to monitoring social distance with live alerts and data sent to a cloud with the user's identity to monitor and process these parameters with geographic location.

In this work, a smart mask that can protect against the COVID-19 spread and also be alert by triggering different alarms in case symptoms appear on the person wearing the mask. Following a brief review of existing masks in the publicly available literature, the proposed mask is discussed in terms of various concepts, sensors, and associated electronics. A prototype is realized, and initial test results are shown and discussed.

2. Related work on smart masks

Several medical, public, and professional masks are available in the market. These masks are intended for different users and protect against viruses, sand particles, and smelly air [6]. The primary categories of masks available in the market related to protection against the COVID-19 virus are discussed. Before and during the coronavirus era, several face-covering masks were proposed. Patents were issued for some designs, whereas others were simply products for study or specific applications. Some designs focused on the public sector, while others were intended for use by medical personnel, specialists, and medical caregivers [7].

A cloth face mask is made exclusively from different types of fabrics. Studies have shown that cloth masks are less effective against SARS-CoV-2 than N95 masks, although they still provide basic protection. They provide the user with protection against airborne contaminants such as pollen and dust particles. However, they have limited suitability for use in the event of a pandemic.

Surgical masks typically consist of three layers: an internal tender absorbent sheet, a central polypropylene obstacle, and an exterior hydrophobic surface. They protect against droplets in a clinical setting and droplet transmission within the range of 1–2 m from an infected person. During the COVID-19 emergency, the WHO recommended medical or surgical masks for all hospital personnel for protection against viral exposure.

Simple masks are surgical masks, which are discarded after one day of use, consisting of three layers. Some of these masks are known as N95 face masks or recent KN95 from Korea. Occasionally, they present choroidal changes after long-term use [8], as they were originally intended for surgeries lasting for a short period.

Rough simple masks cover half of the face without measuring parameters but only filtering the breathing air have been manufactured by 3 M, [9] as well as those with carbon filters [10].

New nanofibre membrane masks with excellent comfortability and anti-pathogen functionality prepared using fluorinated carbon nanofibres/carbon fiber (F-CNFs/CF) were reported to replace existing masks [10].

NIOSH and other respirators are seal-tested respirators that are suitable for medical care personnel, especially those who interact directly with patients. This assembly causes congestion around the nose and mouth and has twisted fibres with channels. Owing to the advent of the Internet of Things (IoT) and integrated miniature devices, numerous sensors have been added to conventional masks to provide them with additional functions, rendering them intelligent. Such masks can monitor the health of medical personnel and may trigger alarms when needed.

Approximately a thousand patents dealing with antiviral masks and ranging from an extremely simple mask with one sensor to multiple sensors for different applications have been reported. A few masks are referenced in these studies [11], [12], [13]. Most of these masks are air filters and cannot provide air quality detection data for breathing, storage of information or post-processing, and collection of data via USB.

Recently, Xiaomi has proposed an e-mask [14] that is not yet on the market. It is equipped with a microprocessor to measure air cleanliness after recording daily data. No alarms have been programmed. For post-treatment, these data can be downloaded through the pulled processor taken from the mask and connected to a PC.

Additionally, some researchers in Singapore have attempted to develop an integrated monitoring system that can easily be fitted to any face mask to monitor the wearer for health indicators associated with COVID-19. The attached sensors measure skin temperature, blood oxygen saturation, blood pressure, and heart rate, all of which are parameters associated with the COVID-19 virus [15]. The protective part of the mask is a substrate made of a polymer material. The overall design integrates all three sensors into one chip. Later, a Bluetooth device will be added to transmit real-time data to a smartphone.

A fully transparent face mask developed at the Birmingham School of Medicine is expected to revolutionize the masking and monitoring of COVID-19 symptoms [16]. The design consists of a clear plastic mouth covering, clear nose barriers, chin barriers, and ergonomic looped arms that fasten around the wearer's ear.

For additional protection, the wearer can attach N95 filter cushions around the mask barrier for use in the medical environment. The arms of the mask are expected to have built-in sensors that monitor temperature and pulse for early detection of viral infections, such as COVID-19. The authors state that the data collected by the mask's sensors could provide barrier protection against the transmission of viral infections such as COVID-19 and could also change the diagnostic process [16].

Engineers at Northwestern University have developed a smart sensor platform for face masks, known as ‘Fitbit for the face’ or ‘Facebit’ [17,18]. A small lightweight sensor with a tiny magnet is attached to an N95, cloth, or surgical face mask. It can measure the user's respiration rate, heart rate, and the length of time the mask is worn, in real-time and potentially replace cumbersome testing by measuring mask fit. This information is wirelessly transmitted to a smartphone app, which contains a dashboard for real-time health monitoring. The app can alert the user immediately to issues such as elevated heart rate or a leak in the mask occurs. The physiological data can also be used to predict fatigue, physical health status, and emotional state.

This review shows that no patents on masks that include temperature sensors to measure human body temperature and sneezing, in addition to monitoring social distance with live alerts and data sent to a cloud with the user's identity to monitor and post-process these parameters with geographic location, have been reported. Our study helps address this gap.

3. Proposed conceptual approach

We propose a protective e-mask made of transparent Perspex with a free face shape that covers the face and also seals the nose and mouth for breathing through a replaceable surgical filter, e.g., N95, while keeping the same plastic mask.

The smart mask (Fig. 1 ) is equipped with a contactless infrared temperature sensor, LED, to show the progressive temperature in the range of 35–38.5 °C. The LED lights up green at temperatures below 36.6 °C, then orange until 37 °C, and red for temperatures exceeding 37 °C. The temperature limits are adjustable. The red color indicates immediate action. An accelerometer sensor is used to detect the rapid movement of the mask and sends data to the microcontroller. A proximity sensor for a one-meter distance (adjustable) provides small alarms for social distancing (optional). The next feature involves connecting to the Internet for user ID (or tag) and their location via Google Maps. All these data, e.g., ID, GPS location, and temperature, are automatically stored per user in a cloud, assuming hot spot/internet is available.

Fig. 1.

Concept of the IoT smart mask.

The data is regularly updated in the cloud for a variety of applications. Users can be informed of their temperature, sneezing, and distancing history with the location. Overall, considerable amounts of data are used for further analysis and tracing to define possible coronavirus contamination and the history of users wearing this smart mask. A general overview of this concept is shown in Fig. 2 and described in the following proposal.

Fig. 2.

Overall progressive planned functions for the facemask.

The proposed smart mask complies with international health requirements, especially for filters. In comparison to existing masks discussed Section 2, this mask show superiority in various aspects mainly in data acquired in the cloud for further analysis including identification of the user, geographic location and all bio parameters values that some of them are common with other proposed masks

Initial prototypes can be produced using 3D printing. However, for mass production injection moulding manufacturing is required, e.g., using a die design. The cost for mass production is expected to be approximately within the range of 8–12 USD per piece. Based on previous investigations and current market needs, as well as the global spread of the COVID-19 virus, we intend to develop a comfortable protective mask that covers the entire face against the COVID-19 pandemic and has a replaceable air filter with remote data storage that includes the following parameters:

• • ○
    ID registration.
  • ○
    Geographic location.
  • ○
    Human body temperature.
  • ○
    Records of coughing/sneezing.
  • ○
    Proximity distance for social distance requirement.
  • ○
    Data are transmitted wirelessly to a cloud for post-processing.

Objectives of the proposed design:

• • ○
    To protect the user from contamination.
  • ○
    Monitor at least one vital sign that usually indicates possible contamination.

Expected outcome:

• ○Fully functional prototype with cloud connectivity.
• ○Alarms set for alerts and feedback.

The e-mask is used as an IoT device for users to connect instantly in an automated way. The basic idea for using the IoT is to continuously collect data from multiple sensors with an almost automatic wireless connection with local data processing, resulting in fast decision-making in real-time. Moreover, if data becomes extensive, it is transferred to a cloud, as will be explained subsequently, but with full monitoring. Although it is currently used for COVID-19 as a topical agent, it can also be used for similar situations, such as sandy wind, long-lasting smoke, and the spread of other types of viruses.

The IoT function is crucial for organizing sub-functions assigned to the mask with connections and data collection. It allows seamless recording of data in the cloud, as users only need to identify and record their ID and let the mask proceeds.

3.1. Mask types and manufacturing

Several concepts have been proposed for masks that would provide respiratory and breathing protection through a sealed zone around the mouth and nose. Such masks must cover the entire face and be transparent. The electronics must be housed in a secure area that is not visible for esthetic reasons, and the mask must have an LED with multiple colours, which are noticeable, in front.

Fig. 3 shows version A with a filter and mask covering the entire face with the electronics under the chin. Occasionally, this is apparent as in version B or C. In version B, the mask is divided into a transparent visor for vision, whereas the other half is not transparent to save material costs. The visor snaps onto the core mask. In version D, the entire mask is transparent and the electronics are in an additional mount added to the mask. The wires connecting the LED and alarm are held through the mask. The filter can slide up from the outside for replacement. The inner shield is made of silicone material, and the outer seal designed to cover the large opening on the outer perimeter of a human face is made of silicone.

Fig. 3.

Proposed concepts for masks (PCB: printed circuit board).

The manufacturing process involves 3D printing for prototyping and testing. Different types of polymers are used, e.g., PLA or plastic filament PETG, because of their printing precision, tolerance, and ability to withstand strong cleaning agents. This material has high temperature resistance and good tolerance to widely used chemicals. The prototype shown in Fig. 4 is printed in three different material colours for the electronics holder: white, transparent, and blue.

Fig. 4.

3D Printed mask prototypes.

3.2. Expected configuration of sensors and microcontroller

In this study, the electronic components, including the microcontroller and the different sensors that are interconnected, are referred to as nodes. Thus, a node consists of a microcontroller with a Wi-Fi connection to connect to the cloud and different sensors to measure different vital signs. The overall scheme for the nodes is shown in Fig. 5 . Vital signs data are measured by different sensors and transmitted to the cloud for post-processing. The sensors measure the body and ambient temperatures and distance between the mask wearer and a nearby person. Finally, an accelerometer measures sudden movements due to sneezing or coughing.

Fig. 5.

Overall strategic configuration of the mask sensors and control.

4. System architecture and IoT environment

4.1. Architecture

The planned functions of the smart mask include introducing user registration, measurement of temperature and social distancing to people in front of the user, as well as any vibrations due to sneezing. All data are either processed locally or sent to the cloud for recording. The fundamental block diagram in Fig. 6 provides an overview of the sensor and microcontroller configurations, whereas the flowchart in Fig. 7 shows the scenario of the system's intended functions and the sensor operating conditions.

Fig. 6.

Functional block diagram.

Fig. 7.

Functional flow chart.

4.2. Component specifications

A typical IoT smart mask node consists of a microcontroller and the sensor characteristics and specifications of the microcontroller and sensors, as described in the following section.

4.2.1. Microcontroller

The primary component of the node is the ESP-07S Wi-Fi module developed by Ai-Thinker Technology [19]. Many others, e.g., the Jetson Nano and LoRa boards, could have been selected based on their various advantages. This was based on a proof-of-concept. An alternative board will be selected later for performance analysis as the smart mask evolves toward higher TRLs. The core processor of ESP-07S is a Tensilica L106 ultra-low-power 32-bit microprocessor clocked at 80/160 MHz. It has a Wi-Fi module, which supports the Wi-Fi standard IEEE802.11 b/g/n protocol. The built-in processing and storage capabilities allow the ESP07 to integrate sensors and other application-specific devices via general-purpose input outputs (GPIOs). The first version of the design was developed. It uses the ESP-8266-01 microcontroller [20] with limited functionalities and sensors. Although this version was functional, it was interrupted. The ESP-07 module has been adopted according to the number of GPIOs which can be used for future improvements. A schematic of the circuit with a temperature sensor is shown in Fig. 8 .

Fig. 8.

ESP8266-07 module.

Although enough GPIOs are available to connect all the necessary peripherals, the circuit was limited to the IR temperature sensor and RGB LED indicators for the initial testing. A new functionality added to this design is the ability to upload the code over the air. This means that the device can be reprogrammed wirelessly, without having to be removed from its environment. This is done in hardware using switch S1 connected to GPIO0 (Fig. 8). On wake-up from sleep mode (SM), GPIO0 is checked by the software and is connected through Wi-Fi to the IDE tool, a new code is loaded remotely. The two figures show the deference between ESP8266 and ESP-07 (Table 1 ).

Table 1.

ESP 07 characteristics.

Feature		Specification
Built-in/ Integrated	Processor	Tensilica L106, low power 32-bit MCU
	ADC	10-bit
	Protection	TR switch, Balun
	RF	LNA, PLL
	Power Management	Regulators, and power management units
	Peripherals	SDIO 2.0, (H) SPI, UART, I2C, I2S, IRDA, PWM, GPIO
Wi-Fi	Protocol Stack	TCP/IP
	Protocol	802.11 b/g/n
	Support Modes	2.4 GHz, WPA/WPA2
	Operation modes	STA/AP/STA+AP
	Output Power	+20 dBm, in 802.11b mode
	Wake up	< 2 ms (and transmit packets)
	Antenna	Supports antenna diversity
Power Consumption	Deep sleep	<10 uA
	Power down leakage	< 5 uA
	Standby power	< 1.0 mW

4.2.2. Temperature sensor

The temperature sensor is the MLX90615 contactless infrared temperature sensor that measures both body and ambient temperature [21]. The sensor has an accuracy of less than 0.2 in the range of the human body temperature (Table 2 ). It is assembled inside the front of the mask to measure the temperature in contactless mode at a distance of a few millimetres from the forehead.

Table 2.

IR contactless temperature sensor characteristics.

Sensor Model	Signal Type	Resolution	Dimensions(mm)
MLX90614/15	2-wire serial SMBus I2C or PWM	0.02 to 0.2 °C	4.1 × 9.15

4.2.3. Proximity sensor

The HCR04 sensor uses ultrasound to determine distances. It has a transmitter unit that generates an ultrasonic wave and a receiver that receives the reflected wave. To compute the distance, a small built-in processor computes the time between when the wave is emitted and the echo is received. This sensor can be miniaturised by selecting alternative smaller prototypes with similar or improved characteristics given in Table 3 where key parameters are quantified.

Table 3.

Ultrasound sensor characteristics.

Feature	Value
Operating Voltage	5 V
Operating Current	15 mA
Operating Frequency	40 kHz
Max Range	4 m
Min Range	2 cm
Ranging Accuracy	3 mm
Measuring Angle	15°
Trigger Input Signal	10 µS TTL pulse
Echo Output Signal	Input TTL level signal range in proportion
Dimension	45 × 20 × 15 mm

4.2.4. Accelerometer

Various sensors are used for the detection and recognition of human motion and gestures [22]. They are specifically used to detect falling, coughing, and sneezing in infected patients. Detecting such parameters helps in monitoring the behavior of elderly people and patients suffering from specific diseases [23,24]. One of the most important sensors used in this context is an accelerometer and a gyroscope [25]. In the current design, an ADXL335 accelerometer is used [26]. It was chosen for its remarkable specifications:3-Axis ±3 g, small footprint (4 mm × 4 mm), and low power consumption (350 μA).

4.2.5. RGB LED indicator and buzzer

Both the RGB LED and buzzer are used as indicators. An RGB LED can produce red, green, and blue lights. Notably, different colours can be produced by configuring the intensity of each LED. These light colours are used for various indicators and alarms. A buzzer is a sound generator that converts an electrical signal into sound. It consists of an electromagnet with a magnetic membrane placed on the housing. When the electromagnet is energised by current, the membrane begins to oscillate, producing an audible signal (Table 4 ). The frequency of the signal can be controlled by software. This frequency is chosen to fall within the normal human audible range (800–1200 Hz).

Table 4.

Passive alarm buzzer.

Feature	Value
Model Number	12,085
external material	plastic
Color	black
Voltage	3 to 12 V
Resistance	16 Ω
Size	12 × 13 mm

4.3. Remote server

Data from sensors are registered and sent to the cloud via Thingspeak, where they are stored in either a private or public channel. By default, Thingspeak stores data in private channels, but the data is openly accessible. Post-processing of data in a Thingspeak channel is possible for operating this project.

The remote server collecting and storing data is Thingspeak (Fig. 11) [3]. Such a server provides free access but has limitations in some functionalities. The primary limitation is the number of data packets that can be received, i.e., one packet every 15 s. However, with proper synchronization and because our nodes tend to switch to sleep mode for 3 to 5 min, a period when no data is transmitted, several nodes can all be connected to the server and transmit packets every 3 to 5 min at 15 s intervals. However, this requires proper synchronization between the transmitting nodes.

Fig. 11.

Thingspeak cloud.

The server provides a graphical representation of the data (Fig. 12). However, for our purposes, a web page written in JavaScript reads the data from the server and presents it in tabular form, as shown in Fig. 13. The table shows the MAC address, temperature of the object, ambient temperature, and distance to the nearest person. The MAC address, labelled as the device ID in the table form in Fig. 13 will be linked to each mask in future designs.

Fig. 12.

Thingspeak data visualization.

Fig. 13.

Local data visualization.

5. Power consumption

The Arduino Uno board consumes approximately 42 mA, assuming that no power is consumed by sensors or other components connected to the GPIO pins. Therefore, the board's power consumption is approximately 40 mA. The onboard voltage regulator has a quiescent current draw of 10 mA even when the processor is in sleep mode. Therefore, even when the sleep mode is enabled with a voltage supply of 3.6 V, the power consumption is approximately 10 mA or 36 mWh. If the system is permanently asleep and powered by a 3.6-V battery delivering approximately 1000 mA, or 3.6 Wh, and all power losses are ignored, such a battery will power the sleeping board for 100 h or a little over 4 d

Table 5 shows the parameters involves in the battery life calculations as well as the related equations. These are used for theoretical calculation of the power consumption. Owing to power limitations as a consequence of using lithium-ion batteries and the fact that the system does not need to constantly send data, the deep-sleep mode of the system cannot reduce power consumption. Therefore, the board is powered directly with a 3.6-V battery, bypassing the onboard regulator. As our nodes enter sleep mode for 3 to 5 min, the period when no data is transmitted, the server is used as the base platform to store our data.

Table 5.

Battery life calculation [4].

Parameter	Unit
C = Capacity rating of the battery	mAh
As = Current of the device when sleeping	mA
Aw = Current of the device when awake	mA
Wph = Number of wakeups per hour	60/5 = 12 to 60/3 = 20
Wt = Duration of a single wake	Ms
c = C*0.85 de-rated capacity	mAh
Twph = Wph*Wt	wake time per hour (ms)
Tsph = msph-Wtph	sleep time per hour (ms)
Aavg = ((Aw*Twph)+(As*Tsph)) / 3,600,000	Current on average (mAh)
hours = c/Aavg	Battery life (h)
days = (c/Aavg)/24	Battery life (d)

The ESP8266 series chip (Fig. 14) provides three configurable sleep modes: mode sleep, light sleep, and deep sleep. Users can choose and configure their sleep mode as required via the software [5]. However, the sleep mode is not enabled by the manufacturer on the ESP8266-01. To enable sleep mode on the module, the board f needs to be modified. The modification consists of connecting GPIO16 (pin 8), which is available on the onboard integrated circuit, to the reset pin of the ESP8266 breakout board. This will connect the output of the onboard RTC to the reset pin to wake up the board after entering sleep mode.

Fig. 14.

ESP8266 main MCU pinout.

The deep-sleep mode is chosen because it has the lowest power consumption. This is done through the software with the command ESP.deepSleep(sleep_time_seconds*factor); where a factor is a value used to convert the time from microseconds to seconds. Table 6 gives a protocol for sleep mode with quantified consumed current. In our case, it is 10⁶. Once the time has elapsed, the device is reset by a signal from the onboard real-time clock circuit, which then wakes up and begins running the code again. By using a similar approach, the estimated runtime of a 1000-mAh battery would be approximately 27.6 h. With average usage of 6 to 8 h per day, such a battery would last between 3 and 5 days.

Table 6.

ESP8266 sleep modes.

Item		Modem-sleep	Light-sleep	Deep-sleep
Wi-Fi		OFF	OFF	OFF
System clock		ON	OFF	OFF
Real-Time Clock (RTC)		ON	ON	ON
CPU		ON	Pending	OFF
Substrate current		15 mA	0.4 mA	∼ 20 μA
Average current	DTIM = 1	16.2 mA	1.8 mA
	DTIM = 3	15.4 mA	0.9 mA
	DTIM = 10	15.2 mA	0.55 mA

The system in Fig. 10 operates according to the scheme shown in Fig. 7. It first checks whether an update is needed or not. Otherwise, it proceeds similarly to the previous design in Fig. 9 . In terms of power consumption and owing to similarity, the design has a consumption that is approximately the same as the previous design in Fig. 9 (ESP07), except for the number of connections and components. Power consumption of the ESP07 based board are given in Table 7 . Power consumption in both sleep and normal mode are shown in Table 8 . These were used in the theoretical power consumption estimation of the current mask board.

Fig. 10.

ESP07-based design electric circuit.

Fig. 9.

ESP8266-based design electric circuit.

Table 7.

ESP-07 based board power consumption.

Component	Normal Mode	Sleep Mode/ Leakage Current
ESP8266–07	200 mA	10 µA
Temperature Sensor: MLX90615	10 mA	3 µA
RGB LED	20 mA	10 µA
Total Current Consumption	230 mA	23 µA

Table 8.

Power consumption of the different devices.

Component	Normal Mode		Sleep Mode
Microcontroller (MCU)	10 mA		< 10 µA
MCU Board	42 mA		10 mA *
ESP8266–01	100–200 mA		10 µA
Temperature Sensor: MLX90615	1.5–10 mA		Sleep Mode: 3 µA, Leakage Current: 0.25 µA
Proximity Sensor: HCR04	Minimum	3.4 mA	-
	Working	15 mA
	Peak	100 mA
RGB LED	20 mA		Leakage Current: 10 µA
Buzzer	30 mA		–

To further reduce power consumption, the two sensors were switched off during sleep mode. This was achieved by connecting the Vcc of the sensors to the digital pins of the microcontroller. The pins were controlled with software. Once the measurements were obtained and data were sent to the server, the sensors were turned off, and the MCU entered the deep-sleep mode. When the MCU returned to normal mode, the sensors were powered by activating a high voltage on their Vcc pins. The sleep mode duration was set between 3 and 5 min. This time was deemed sufficient to prevent serious temperature fluctuations. The node got up again, sensors were switched on, and time was given to perform calibration. Measurements were taken, and the cycle was repeated. Battery life was estimated using the approach described in [27]. As shown in Table 9 , the battery was expected to last for approximately 19.4 h. Moreover, with an estimated usage of 6–8 h per day, the battery could last for 2–3 d. However, further reductions in power consumption were sought by effectively using the deep sleep mode of the different components, which increased the battery life. The supplied power could be used via radio frequencies to energize the electronics or preferably to charge the batteries, considering the power management in a WSN format, as performed in [28,29]. A comparison between theoretical and experimental power consumption of the whole mask power is shown in Table 9. It shows a difference in battery lifetime between experimental and theoretical calculation based on the parameters described earlier. If sleep mode (SM) is added, the battery can save up to 1 h, i.e., 6.25% of the total life time.

Table 9.

Total power consumption of the smart mask.

	Theoretical battery lifetime	Experimental battery lifetime without sleep mode	Experimental battery lifetime with sleep mode (12 times SM) (SM=2–3 min)
Battery lifetime	19.4 h (∼2–3 days)	16.1–16.5 h (< 2–3 days)	17.1 h(< 2–3 days)
	Theoretical usage time Per day	Estimated usage time per day	Effective usage time Per day
Mask usage	7–9 h	6–8 h	6.2–8.1 h

6. Experiments and results

Several tests are performed before assembly. The infrared temperature sensor can measure the body temperature but is limited by the battery lifetime estimated at 3 d. Because measurements are scheduled every minute in an extensive set of tests, the system goes into a deep sleep and wakes up the next minute to measure. Later, the cycle of measurement can be scheduled to occur every hour, and in this case, battery life can be extended to last more than a month. Power consumption can be estimated by measurements with several existing sensors and a microcontroller to calculate the battery life. The process of powering up the smart mask, connecting to the Internet, and registering the user's ID was checked for its systematics. Some errors were identified that are primarily related to connectivity issues with internet providers [30,31].

Several complete masks were equipped with sensors and tested for a few days. More than a dozen participants agreed to test the masks. Fig. 15 shows an excerpt from the large data received in the cloud, showing the behavior of two masks in terms of temperature and social distancing over one day. The users could be easily identified as each user had a specific identity registered by the microprocessor. The cloud used for the masks was registered for testing at thingspeak.com, where post-processing of data can be done using MATLAB (Fig. 16 ).

Fig. 15.

Measurement results of a sample of two masks.

Fig. 16.

Visualization of record and storage of measurement of one mask.

When the temperature exceeded 37 °C, it can be seen in the graph inside the red zone. Similarly, when social distancing is less than 2 m, the two users can be detected, and their geographical location identified.

We did not observe any congestion of information in the cloud owing to the small number of masks used. However, this can be solved with multiple accounts in the cloud to which masks are assigned. The location of each mask can be determined using Google location, as shown in Fig. 16. This figure also shows instant measurement records in the cloud.

The wearer is being monitored in terms his bio-parameters data sent on regular basis. He is informed about himself and his surrounding for protective information, but also data is recorded for history behavior to determine whether a possible contamination has happened and who are the users involved for their protection.

Further, the mask is reusable, as filters can be changed regularly. However, it can also be used as a disposable mask if it is to be used by individual users over a longer period. It is difficult to add a sensor to sensor to inspect the mask cleanness for the time but it remains optional if needed.

Since data is transferred through internet and carries personal data including identification, it is requested that these data are protected or encrypted. The conditions of use of the mask and features to be extended and may vary depending on the environment whether it is dry without humidity, e.g., indoors, or with humidity, sand storm etc. in outdoors.

7. Conclusion

The demand for respiratory protective equipment is growing in the market. Any additional smart feature becomes standard as it adds value to the consumer.

Given the current international agreement to wear masks in public areas that provide filtered breathing during the day and reduce the likelihood of spreading or contracting viruses, this study presented a smart mask with an ecosystem and intelligent features that can protect breathing through filters, monitor body temperature, sneezing activity, and distanciation condition, and record the geographic location of the mask wearer after ensuring an ID registration.

Data could be processed locally to trigger alarms related to temperature and social distancing. The remaining data was sent to a cloud for post-processing to record the history of the mask wearer, including the geographic dynamic location, to track the possible spread of contamination. The prototype and measurement results presented demonstrated the practicality and potential utility of mass numbers. This smart mask can be a platform for future development and integration of additional functions.

The superiority of this mask comes also from the data post processing with warnings where the wearer is being monitored in terms his bio-parameters data sent on regular basis. He is informed about himself and his surrounding for protective information, but also data is recorded for history behavior to determine whether a possible contamination has happened and who are the people involved. Data is collected in the cloud from several users. The system with post processing has a better mapping of the users safe from those who are sick. Information may warn the safe users to get away or warn those sick to keep a assistance as a first measure.

Data collected from users in a cloud system are of considerable interest in many aspects, including prevention and healthcare decision-making. The current known limitations of the system are Wi-Fi connectivity and cloud capacity, which depend on the size of the data collected and post-processed.

Artificial intelligence can easily be added depending on the user's situation and condition. The mask has been patented in the US Patent Office under the reference US Patent 2022/0208391 A1 [32]. This smart mask can be reused, and the cost of mass production does not exceed 10 USD.

Declaration of Competing Interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Samir Mekid reports financial support was provided by KFUPM University. Samir mekid reports a relationship with King Fahd University of Petroleum & Minerals that includes: employment and funding grants. Samir mekid has patent smart mask pending to USPatent. N/A.

Data availability

• The data that has been used is confidential.