Article

Smart Sensing Strip Using Monolithically Integrated Flexible Flow Sensor for Noninvasively Monitoring Respiratory Flow Peng Jiang, Shuai Zhao and Rong Zhu * Received: 19 October 2015; Accepted: 10 December 2015; Published: 15 December 2015 Academic Editor: Steffen Leonhardt State Key Laboratory of Precision Measurement Technology and Instrument, Department of Precision Instrument, Tsinghua University, Beijing 100084, China; [email protected] (P.J.); [email protected] (S.Z.) * Correspondence: [email protected]; Tel.: +86-10-6278-8935

Abstract: This paper presents a smart sensing strip for noninvasively monitoring respiratory flow in real time. The monitoring system comprises a monolithically-integrated flexible hot-film flow sensor adhered on a molded flexible silicone case, where a miniaturized conditioning circuit with a Bluetooth4.0 LE module are packaged, and a personal mobile device that wirelessly acquires respiratory data transmitted from the flow sensor, executes extraction of vital signs, and performs medical diagnosis. The system serves as a wearable device to monitor comprehensive respiratory flow while avoiding use of uncomfortable nasal cannula. The respiratory sensor is a flexible flow sensor monolithically integrating four elements of a Wheatstone bridge on single chip, including a hot-film resistor, a temperature-compensating resistor, and two balancing resistors. The monitor takes merits of small size, light weight, easy operation, and low power consumption. Experiments were conducted to verify the feasibility and effectiveness of monitoring and diagnosing respiratory diseases using the proposed system. Keywords: respiratory disease; wireless monitoring; micro sensor; flexible hot-film flow sensor; wearable device

1. Introduction According to the report of World Health Organization (WHO), respiratory diseases have become one of the top major killers in the past decade [1]. Apnea Syndrome, Asthma, and Chronic Obstructive Pulmonary Disease (COPD) are the most common respiratory diseases that bring serious risks to human health, and even death. Sleep apnea is an involuntary cessation of breathing that occurs while the patient is asleep [2]. Obstructive sleep apnea syndrome (OSAS) directly relates to human respiration, which is suffered by 20%–40% of elderly people, and more than half patients remain undiagnosed [3]. Risk factors include being male, overweight, and over the age of 40. As a matter of fact, however, sleep apnea may strike anyone at any age, even children [4]. Most of respiratory diseases are chronic in nature and have a major impact not only on the individuals with diseases, but also on community, which bring about significant public health burdens. Therefore, effective methods of early detection and diagnosis are significantly necessary [5]. Respiratory airflow monitoring is vital to diagnose OSAS and other respiration diseases. In order to diagnose OSAS, a system called polysomnography (PSG) [6] is commonly used in sleep labs and operated by physicians. PSG is the current preferred diagnostic modality but is relatively inconvenient, expensive, and has low efficiency [7]. Consequently, a portable monitor (PM) has been proposed as a substitute for PSG in pre-diagnostic assessment of patients with suspected

Sensors 2015, 15, 31738–31750; doi:10.3390/s151229881

www.mdpi.com/journal/sensors

Sensors 2015, 15, 31738–31750

OSA [8]. PM requires less technical expertise, is less labor intensive and time consuming, and is easier for patients to access [9,10]. Our group previously reported a wireless portable system for monitoring respiratory diseases. We used a micro thermal flow sensor to monitor respiratory airflow, a tri-axis micro accelerometer to monitor body posture, and a microphotoelectric sensor to monitor blood oxygen saturation [11]. Respiratory symptoms to be monitored including, but not limited to, effort, airflow, snoring, end-tidal CO2 , esophageal pressure, breathing humidity. There are several kinds of respiratory sensors that have been reported. One is indirectly monitoring respiration from variation in thorax volume, the other is using airflow sensor to directly detect respiratory airflow [12]. The former method includes transthoracic impedance sensors, micro and macro bending fiber-optic sensors [13], 3D image sensors, and ultrasonic sensors, which could detect the variation of chest or abdominal circumference. However the indirect methods are sensitive to body movements, which may result in invalid measurements. Respiratory airflow is a key vital signal for monitoring respiratory disorder. The apnea and hypopnea index are more relevant with diagnosis of OSAS. Several kinds of airflow sensors are used to monitor respiratory airflow, such as pressure sensor, hot-wire/hot-film sensor, infrared thermography [14], and ultrasonic flowmeter [15]. The differential pressure flow sensor and the hot-wire/hot-film flow sensor are the most common flow sensor used in the portable medical instruments and some wearable devices. Differential pressure flow sensors generally need a pipeline and diaphragm structure. In recent years, both Fleisch pneumotahographs and variable orifice meters have a linear response; hence, they overcome the hurdle related to the low sensitivity at low flow rate of the fixed orifice meters [16,17]. The intensity-modulated fiber-optic sensors can monitor the humidity of exhaled airflow. The condensed humidity substantially alters the interaction between the light from the optical fiber and the surrounding medium, which can be monitored by a photo detector [13]. The IRT method can remotely sense the breathing dynamics by detecting the temperature variation of the exhale and inhale air [14]. Comparatively, micro hot-wire/hot-film sensors have a small size, simple structure, low cost, high sensitivity, and capability of integrating with on-chip circuitry [18]. For monitoring respiratory airflow, most monitors use nasal cannulas which have two prominent small pipes inserted into the nostrils to conduct nasal airflow to the flow sensors. The narrow tube may hinder nasal airflow and increase the respiratory resistance, thereby results in unsmooth and uncomfortable breathing for the patient. Furthermore, the long nasal cannula is easily squeezed by the human body and twined with limbs. Effective respiratory airflow monitor with noninvasive and comfortable detection is increasingly imperative. In this paper, we propose a miniaturized flexible respiration monitor (termed smart strip) that can be attached on human upper lip underneath the nostril to implement non-invasive, wireless, and real-time respiratory flow monitoring without use of nasal cannula and mask. The device utilizes a tailor-designed micro flow sensor, which is monolithically-integrated with a temperature-compensating resistor and two balancing resistors on single chip. The monitoring system comprises a flexible flow sensor attached on a molded flexible silicone case, a miniaturized conditioning circuit with a low power Bluetooth4.0 LE module packaged in the case, and a personal mobile device that wirelessly acquires sensor data, executes extraction of vital signs, and performs medical diagnosis. 2. System Design 2.1. System Overview Figure 1 shows a flexible respiration monitor structured like a strip, which is comprised of a micro hot-film flow sensor, a signal conditioning circuit, a Bluetooth4.0 LE module, and rechargeable battery. The monitor is wirelessly connected with a personal computer (PC) or a smart phone via Bluetooth communication. The monitor (also termed smart strip of respiration monitor, SSRM) is flexible that can be attached on the upper lip underneath the nostrils. The respiratory flow breathed

31739

Sensors 2015, 15, 31738–31750

Sensors 2015, 15, page–page

out from the nostrils blows over the strip and is detected in real-time by the hot-film flow sensor attached the The strip. Tube-free configuration makes the respiration the sensorondata. Bluetooth4.0 LE module containing a 12 bit ADC, amonitoring MCU part,non-invasive a GPIO part, and and unobstructed to15,the breath ofisthe patient. Figure 2 shows data the signal flow phone of the system. the SSRM, Sensors 2015, page–page Bluetooth4.0 LE transceiver used to submit respiration to a smart or a PC. In Smart phone the hot-film flow is used to detect the respiratory flow, and the conditioning circuit acquires or PC receives thesensor respiration data, implements medical analysis, records, and displays the results. the sensor Bluetooth4.0 modulecontaining containing aa12 a MCU part, a GPIO part,part, and and the sensor data.data. TheThe Bluetooth4.0 LELEmodule 12bit bitADC, ADC, a MCU part, a GPIO Bluetooth4.0 LE transceiver is used to submit respiration data to a smart phone or a PC. Smart phone Bluetooth4.0 LE transceiver is used to submit respiration data to a smart phone or a PC. Smart phone or PC receives the respiration data, implements medical analysis, records, and displays the results. or PC receives the respiration data, implements medical analysis, records, and displays the results. Respi system

Respiration system

Respi system

apnea: xx hypopnea: xx

Respiration system apnea: xx hypopnea: xx

Conditioning circuit Silicone cover circuit Conditioning Silicone cover Battery

Hot-film flow sensor

Battery

Hot-film flow sensor

Bluetooth Bluetooth

Silicone pedestal Silicone pedestal

Schematic of the system. Figure Figure1.1. Schematic of the system.

Smart SmartStrip Strip Respiration RespirationMonitor Monitor Hot-film Hot-film Flow Flow Sensor Sensor

Conditioning Conditioning Circuit Circuit Control Power

Control Power Supply Supply

Bluetooth4.0 BLE Bluetooth4.0 BLE SOCmodule module SOC ADC, ADC,Data Data Process I I Process

BLE BLE Transceiver

Transceiver

Smart Phone or PC with Bluetooth4.0 Smart Phone ormodule PC with Bluetooth4.0 module BLE Data Transfer Transceiver

To Internet

BLE Transceiver

Data Transfer To Internet

Data Process II And Recorder

Result Display

Data Process II And Recorder

Result Display

Figure 2. The block diagram of signal flow.

Figure 2. 2. The The block block diagram diagram of of signal flow. flow. Figure 2.2. Design and Fabrication of Monolithically Integrated Flexible signal Hot-Film Flow Sensor

anemometers have been widely used in industry, automobile, medical, aerospace, and 2.2. DesignHot-film and Fabrication Fabrication of 2.2. Design and of Monolithically Monolithically Integrated Integrated Flexible Flexible Hot-Film Hot-Film Flow Flow Sensor Sensor other fields. The micro-machined thermal flow sensor based on silicon technology was firstly

Hot-film anemometers havethe been widelyused used in industry, automobile, medical, aerospace, developedanemometers in 1974 [19]. With development of flexible electronics, several medical, flexible flow sensors Hot-film have been widely in industry, automobile, aerospace, and and other fields. The micro-machined thermal flow sensor based on silicon technology was firstly have been reported in recent years [20]. Petropoulos A. et al. developed a flexible gas flow sensor other fields. The micro-machined thermal flow sensor based on silicon technology was firstly which was fabricated on a flexible PCB substrateof and was able to quantifyseveral 0–10 SLPM flow flow rate [21]. developed in [19]. With the development development of flexible electronics, several flexible flow sensors developed in 1974 1974 [19]. With the flexible electronics, flexible sensors Sturm H. et al., developed a thermoelectric mass flow-rate sensor on a 10 μm thick polyimide foil [22]. have been reported in recent years [20]. Petropoulos A. et al. developed a flexible gas flow have been reported in recent years [20]. Petropoulos A. et al. developed a flexible gas flow sensor sensor Bowman B. et al., developed a respiration sensor containing thermoresistive elements, which were which quantify 0–10 0–10 SLPM which was was fabricated fabricated on on aa flexible flexible PCB PCB substrate substrate and and was was able able to to quantify SLPM flow flow rate rate [21]. [21]. screened on the substrate using conductive ink [23]. Sturm developed a a thermoelectric thermoelectric mass µm thick Sturm H. H. et et al., al., developed mass flow-rate flow-rate sensor sensor on on aa 10 10 μm thick polyimide polyimide foil foil [22]. [22].

Bowman B. et al., developed a respiration sensor 3containing thermoresistive elements, which were screened on the substrate using conductive ink31740 [23]. 3

Sensors 2015, 15, 31738–31750

Bowman B.15,etpage–page al., developed a respiration sensor containing thermoresistive elements, which were Sensors 2015, screened on the substrate using conductive ink [23]. The integrated integrated hot-film hot-film sensor sensor isis shown shown in in Figure Figure 3, 3, combines combines four four resistors resistors on on one one chip, chip, The including aa hot-film hot-film resistor resistorRh, Rh, aa temperature temperaturecompensating compensatingresistor resistorRc, Rc,and and two two balancing balancing resistors resistors including Ra and and Rb, Rb, all all of of which which serve serve as as legs legs of of aa Wheatstone Wheatstone bridge bridge[24]. [24]. The The fabrication fabrication of of the the sensor sensor is is Ra shown in Figure 3c. The sensor is fabricated on a flexible polyimide substrate by incorporating the shown in Figure 3c. The sensor is fabricated on a flexible polyimide substrate by incorporating the printed circuit circuittechnique techniquewith withthe the micromachining micromachiningsputter sputtertechnique. technique. The The hot-film hot-film flow flow resistors resistors and and printed temperature-compensating resistors are made of platinum due to its high temperature coefficient of temperature-compensating resistors are made of platinum due to its high temperature coefficient of resistance (TCR). (TCR). The The two two balancing balancing resistors resistorsas asfixed fixedresistors resistorsare aremade madeofofNichrome Nichrome(80/20 (80/20 wt%), wt%), resistance namely Ni80Cr20, Ni80Cr20, due due to to its its low low thermal thermal sensitivity sensitivity and and low low conductivity. conductivity. Figure Figure 3d 3d demonstrates demonstrates namely the prototype prototype of of the the fabricated fabricated sensor. sensor. The the hot-film hot-film flow flow resistor resistor and and the the the The average average TCR TCR of of the temperature-compensating resistor resistor isis about about 1400 1400 ppm/K ppm/K and and the the average average TCR TCR of of two two balancing balancing temperature-compensating resistorsisisaround around70 70ppm/K. ppm/K. resistors

Figure Figure 3. 3. (a) CTD circuit circuit of of the the flow flow sensor; sensor; (b) (b) the the design design of of the the monolithically monolithically integrated integrated hot-film hot-film flow flowsensor; sensor;(c) (c)the thefabrication fabricationprocesses processesof ofthe thesensor; sensor;and and(d) (d)fabricated fabricatedsensors. sensors.

Though many many micro micro hot-film hot-film flow flow sensors sensors on on silicon silicon or or polyimide polyimide substrate substrate have have been been Though realized [19], monolithic integration of the flow sensor with its conditioning circuit, especially realized [19], monolithic integration of the flow sensor with its conditioning circuit, especiallyonona airflow, aflexible flexiblesubstrate substrateisisstill stillproblematic. problematic.As Asmentioned mentionedabove, above,to to directly directly detect detect the the respiratory respiratory airflow, an effective airflow sensor is important. In traditional methods, the airflow sensors are usually an effective airflow sensor is important. In traditional methods, the airflow sensors are usually embedded into a tube to monitor the respiratory flow blowing through it. The respiratory is embedded into a tube to monitor the respiratory flow blowing through it. The respiratoryflow flow collected and pipelined byby a nasal tube or or nasal mask, which unavoidably hinders breathing. To is collected and pipelined a nasal tube nasal mask, which unavoidably hinders breathing. overcome the problem, the proposed monitor adopts a monolithically integrated flexible hot-film To overcome the problem, the proposed monitor adopts a monolithically integrated flexible hot-film flow sensor sensor attached attached on on aa flexible flexible silicone silicone case case which which can can be be used used to to monitor monitor comprehensive comprehensive flow respiratory airflow in a noninvasive way. respiratory airflow in a noninvasive way. The flow flow sensor sensor utilizes utilizes the the hot-film hot-film resistor resistor Rh Rh as as both both Joule Joule heater heater and and temperature temperature detector. detector. The Under aa constant constant bias Under bias power power and and zero zero flow flow rate, rate, the thehot-film hot-filmresistor resistorisisheating heatingand andachieves achievesa steady-state temperature, which means the heat transfer system reaches equilibrium. If an a steady-state temperature, which means the heat transfer system reaches equilibrium.external If an airflow airflow passes, passes, the hot-film resistor experiences forced convective the external the hot-film resistor experiences forced convectivecooling. cooling. Accordingly, Accordingly, the temperature of of the the hot-film hot-film resistor resistor decreases decreases and and makes makes the the resistance resistance changes, changes, and and thus thus provides provides temperature the information of the flow rate that dominates the cooling rate. A constant temperature difference the information of the flow rate that dominates the cooling rate. A constant temperature difference (CTD) mode mode [25] [25] shown shown in in Figure Figure 3a 3a is is used used to to operate operate the the hot-film hot-film sensor sensor due due to to its its advantages advantages of of (CTD) highsensitivity sensitivityand andfast fastresponse, response,where whereRhRh hot-film resistor with about 90atΩ23 at˝23 °C,isRc is high is is thethe hot-film resistor with about 90 Ω C, Rc the the temperature-compensating resistor with about 474 Ω, Rb and Ra are the balance resisters with temperature-compensating resistor with about 474 Ω, Rb and Ra are the balance resisters with 50 Ω 50 Ω263 and Rtadjust is an adjust resistor for modulating the overheat ratio of 0.5%. and Ω,263 Rt Ω, is an resistor for modulating the overheat ratio of 0.5%. To make the monitor flexible and bendable so as to be capable to attach on on thethe human upper lip, To make the monitor flexible and bendable so as to be capable to attach human upper the flexible hot-film flow sensor was adhered on the outer surface of a molded flexible case. The lip, the flexible hot-film flow sensor was adhered on the outer surface of a molded flexible case. packaging case case of theofmonitor was molded by using (PDMS), which is flexible The packaging the monitor was molded by polydimethylsiloxane using polydimethylsiloxane (PDMS), which is and biocompatible. The conditioning circuit of the hot-film sensor, the Bluetooth module, and the flexible and biocompatible. The conditioning circuit of the hot-film sensor, the Bluetooth module, battery were distributed and packaged in the case. Figure 4 shows the and the battery were distributed and packaged in flexible the flexible case. Figure 4 showsa aprototype prototype of of the developed SSRM SSRMwith withaasize sizeof of55mm mmˆ × 25 developed 25 mm mm × ˆ100 100mm. mm. 31741

4

Sensors 2015, 15, 31738–31750 Sensors 2015, 15, page–page

Sensors 2015, 15, page–page

Figure 4. The prototype of developed monitor and demonstration of attaching it on the upper lip.

Figure 4. The prototype of developed monitor and demonstration of attaching it on the upper lip.

2.3. Low Power Consumption Design

2.3. Low Power Consumption Design

Power consumption is an important issue to be considered in system design. Low-power

Figure 4. The prototype developed of amplifier attaching it on the upper Power consumption is anofsuch important issueand to demonstration beoperational considered in system design. components were selected, as amonitor low-power which has lip. a Low-power power consumption less than 100 μA at The power operational consumption of the Wheatstone components wereof selected, such as3.3 a V. low-power amplifier which bridge has awas power 2.3. Low Power Consumption Design reduced by a low100 ratio of at Rb/Rh (around 0.5) andconsumption using an appropriate overheat ratiobridge (0.5%) was consumption ofsetting less than µA 3.3 V. The power of the Wheatstone consumption an important issue topower be system Low-power to make a tradeoff between the sensitivity and the consumption for thedesign. sensor. The average reduced byPower setting a low ratio isof Rb/Rh (around 0.5) andconsidered using an in appropriate overheat ratio (0.5%) components were selected, such as a low-power operational amplifier which has a and power working power of the conditioning circuit of the hot-film sensor reached about 33 mW the to make a tradeoff between the sensitivity and the power consumption for the sensor. The average consumption of less than 100 μAwas at 3.3 V. The consumption of the aWheatstone monitoring system consumption 43 mW. Topower further save the energy, sleep-modebridge circuitwas was working power of the aconditioning circuit of the sensor reachedoverheat about 33 and the reduced byMCU setting low ratio of Rb/Rh (around 0.5)hot-film andthe using an appropriate ratiomW (0.5%) used. When received a “run” command, it opened MOSFET to power the conditional circuit. monitoring system consumption was 43 mW. further save the to energy, a sleep-mode circuit was tradeoff between sensitivity andTothe power consumption for the The average If to themake MCUa received a “stop” the command, it shut down the MOSFET make thesensor. conditioning circuit used.dormancy, When MCU received a “run” command, it opened the MOSFET to power the conditional working power of the conditioning circuit of the hot-film sensor reached about 33 mW and the and then the MCU entered into sleep mode. Bluetooth Low Energy (BLE) is an emerging monitoring system consumption 43 mW. To save the energy, a sleep-mode circuit was circuit. If the MCU received a “stop”was command, it further shut down thecommunication MOSFET to make conditioning low-power wireless technology developed for short-range and the monitoring used. When MCU received a “run” command, it opened the MOSFET to power the conditional circuit. circuit dormancy,that and then thetoMCU entered into mode. Bluetooth Lowfew Energy applications is expected be incorporated intosleep billions of devices in the next years (BLE) [26]. Inis an If thetoMCU received a “stop” command, it shut down the MOSFET to make the circuit order minimize power consumption, our monitor combined a Bluetooth4.0 LEconditioning moduleand (DA14580) emerging low-power wireless technology developed for short-range communication monitoring dormancy, and then the MCU entered sleep mode. Bluetooth Low Energy (BLE) is an emerging with an MCU to expected construct atodigital datainto acquisition system. Theofaverage working of the BLE [26]. applications that is be incorporated into billions devices in the power next few years low-power wireless technology developed for short-range communication and monitoring module is only about 5 mW [27]. Use the MCU to control the conditioning circuit power supply In order to minimize power to consumption, a next Bluetooth4.0 LE and module applications that is expected be incorporatedour intomonitor billions ofcombined devices in the few years [26]. In acquire the respiration data. Figure 5 shows a flow chart of the monitor. All techniques mentioned (DA14580) with an MCU to construct a digital data acquisition system. The average working order to minimize power consumption, our monitor combined a Bluetooth4.0 LE module (DA14580)power above saved power as much as possible to extend the battery life. In addition, a wireless battery with module an MCU to a digital data acquisition Thecontrol averagethe working power of circuit the BLEpower of the BLE is construct only about 5 mW [27]. Use thesystem. MCU to conditioning charging chip with a low power consumption of 2 μA at 4.2 V was integrated within the SSRM to module is only the about 5 mW [27].data. Use the MCU 5 toshows control athe conditioning circuit power supply and supply and acquire respiration Figure flow chart of the monitor. All techniques realize a wireless charging for the device. acquire the respiration data.as Figure 5 shows a flowto chart of thethe monitor. Alllife. techniques mentioned mentioned above saved power much as possible extend battery In addition, a wireless above saved power as much as possible to MCU extend the battery life. In addition, a wireless battery and BLE battery charging chip with a low power consumption of 2 µA at 4.2 V was integrated within the SSRM charging chip with a low power consumptionInitialize of 2 μA at 4.2 V was integrated within the SSRM to to realize a wireless charging for the device. realize a wireless charging for the device. MCU and BLE SleepMode MCU and BLE Initialize BLE Advertising MCU and BLE SleepMode BLE Connected? BLE Advertising

BLE Disconnected

Yes

BLE BLEUP Wake Connected?

BLE Disconnected No

No

BLE Receive Yes Run Command

MCUBLE Wake Up Wake UP Conditioning Circuit Power On BLE Receive Run

BLE Connected?

MCU Sleep

MCU Conditioning Sleep Circuit Power Off

Command

BLE Receive MCU Wake Up Command Conditioning Circuit Power On Run

Yes

Stop Conditioning Command

Circuit Power Off

Command

BLE Connected?

Yes

BLE Receive MCU Acquires Command Respiration Data Run Send to Phone

Stop Command

Command

MCU Acquires Figure 5. A flow chart of the SSRM.

Respiration Data Send to5Phone

Figure 5. A flow chart of the SSRM.

Figure 5. A flow chart of the SSRM. 5

31742

Sensors 2015, 15, 31738–31750 Sensors 2015, 15, page–page

2.4. 2.4. Algorithm Algorithm for for Extracting Extracting Respiratory Respiratory Parameters Parameters An An algorithm algorithm of of extracting extracting the the respiratory respiratory parameters parameters for for analyzing analyzing and and diagnosing diagnosing respiratory respiratory diseases was designed. Minute ventilation (MV), peak inspiratory flow (PIF), respiratory diseases was designed. Minute ventilation (MV), peak inspiratory flow (PIF), respiratory rate rate (RR), (RR), and and tidal tidal volume volume (TV) (TV) are are the the key key parameters parameters of of respiration. respiration. Respiratory Respiratory minute minute volume volume (or (or minute minute ventilation ventilation or or expired expired minute minute volume) volume) is is the the volume volume of of gas gas inhaled inhaled (inhaled (inhaled minute minute volume) volume) or or exhaled (exhaled minute volume) from a person’s lung per minute [28]. Digital integration of exhaled (exhaled minute volume) from a person’s lung per minute [28]. integration of acquired and thethe maximum value of of respiratory peak in acquiredrespiratory respiratoryflow flowvalue valueininone oneminute minutegets getsMV, MV, and maximum value respiratory peak one minute is defined as PIF. Counting the peaks of respiration waveform in one minute gets the value in one minute is defined as PIF. Counting the peaks of respiration waveform in one minute gets the of RR. of Dividing the minute ventilation by the by respiratory rate obtains the value TV. Table shows1 value RR. Dividing the minute ventilation the respiratory rate obtains theof value of TV.1Table the respiratory parameters calculated from an from experimental data where a health wasadult testedwas in shows the respiratory parameters calculated an experimental data where adult a health normal breathing. Table 1 also list typical normal human respiratory parameters as comparison. tested in normal breathing. Table 1 also list typical normal human respiratory parameters as

comparison. Table 1. Tested respiratory parameters. Parameters

Parameters Tested Value Tested Value Value Normal Normal Value

Table 1. Tested respiratory parameters. MV (L) TV (L) PIF (L/min)

MV (L) 6.1 6–86.1 6–8

TV (L) 0.3 0.3 0.3–0.8 0.3–0.8

PIF (L/min) 21 21 N/A N/A

RR (min´1 )

RR (min−1) 20 20 12–24 12–24

3. Experimental Results and Discussions 3. Experimental Results and Discussions 3.1. Simulation of Human Exhalation and Inhalation Flow Field 3.1. Simulation of Human Exhalation and Inhalation Flow Field In order to theoretically verify the feasibility of the respiratory monitor, a computational order to(CFD) theoretically verify feasibility of theairflow respiratory a computational fluid In dynamic simulation of the human respiration was monitor, conducted to demonstratefluid the dynamic (CFD) of human respiration airflow was conducted demonstrate the ventilation and thesimulation flow distribution. We chose a k-ε RANS turbulent model [29] to to calculate the flow ventilation and the flow distribution. We chose a k-ε RANS turbulent model [29] to calculate the flow field. The simulated exhalation and inhalation processes are shown in Figures 6 and 7 respectively. field. simulated exhalation andwas inhalation processes shownwhich in Figures 6 and 7 respectively. In the The exhalation process, the inlet inside the nostril are through an expiration flow rate of the blew exhalation process, the inlet of was inside the nostril through which expiration flow rate of 3Inm/s out. The flow velocity 3 m/s was determined according to an a typical respiration flow 3 m/s blew out. The flow velocity of 3 m/s was determined according to a typical respiration flow rate rate 18 L/min. The diameter of the nostril was set as 8 mm corresponding to the anthropometric 18 L/min. diameter the nostrilprocess, was set as 8 mm corresponding to the of be an size of an The adult. In theofinhalation the pressure of the nostril asanthropometric an outlet was size set to In the inhalation20process, thea pressure the nostril pressure as an outlet was set to be a pressure aadult. pressure subtracting Pa from standard of atmospheric to simulate a normal inhale subtracting 20 Pa from a standard atmospheric pressure to simulate a normal inhale state. state. The negative pressure of 20 Pa was estimated according to the inhale flow rate. In bothThe of negative pressure of 20 Pathe wasrespiratory estimated flow according to the inhale flow rate. In both of exhalation and exhalation and inhalation, blew through the upper lip, was detected by the SSRM inhalation, the respiratory flow blew through the upper lip, was detected by the SSRM attached there. attached there.

Figure 6. 6. Simulation Simulation of of human human exhalation. exhalation. Figure

31743 6

Sensors 2015, 15, 31738–31750 Sensors 2015, 15, page–page Sensors 2015, 15, page–page

Figure Figure 7. 7. Simulation Simulation of of human human inhalation. inhalation. Figure 7. Simulation of human inhalation.

3.2. 3.2. Calibration and Dynamic Response Response of of the the Airflow Airflow Sensor Sensor 3.2. Calibration and Dynamic Response of the Airflow Sensor A A calibration calibration experiment experiment was conducted conducted to obtain the relationship between the respiratory flow flow A calibration experiment was conducted to obtain the relationship between the respiratory flow and the output signal of the sensor (i.e., the top voltage of the Wheatstone bridge). The experimental and the output signal of the sensor (i.e., the top voltage of the Wheatstone bridge). and for the calibration output signal the sensor (i.e., the of theSSRM Wheatstone bridge). on Thethe experimental setup is shown in 8. was attached upper lip of setup for calibration is of shown in Figure Figure 8. top Thevoltage developed setup for calibration is shown in Figure 8. The developed SSRM was attached on the upper lip of human model.The Theoutlet outlet a pipe with an inner diameter of 9was mmconnected was connected into the human face model. of of a pipe with an inner diameter of 9 mm into the nostril human face model. The outlet of a pipe with an inner diameter of 9 mm was connected into the nostril nostril of a face model, by which a controlled airflow was breathed out the nostril and blew through of a face model, by which a controlled airflow was breathed out the nostril and blew through the of a face model, by which a controlled airflow was breathed out the nostril and blew through the the SSRM. inlet tube was connectedtotoananairflow airflowmeasurement measurementand and control control system (Fluke SSRM. TheThe inlet of of thethe tube was connected (Fluke SSRM. The inlet of the tube was connected to an airflow measurement and control system (Fluke MOLBOX1+) that accurately adjusted the airflow rate. Figure 9a illustrates the relationship between between MOLBOX1+) that accurately adjusted the airflow rate. Figure 9a illustrates the relationship between the airflow and the output voltage ofofthe sensor. The ofof thethe sensor outputs is estimated to thethe airflow and Theuncertainty uncertainty sensor outputs estimated airflow andthe theoutput outputvoltage voltage of the the sensor. sensor. The uncertainty of the sensor outputs is is estimated be about ˘5% and the bias stability of the sensor is about 5 mV as shown in Figure 9b. According to to to be about ±5% and the bias stability of the sensor is about 5 mV as shown in Figure 9b. According be about ±5% and the bias stability of the sensor is about 5 mV as shown in Figure 9b. According the King’s law, thethe relationship between the and the sensor output voltage to to the King’s law, relationship between the airflow(denoted (denotedas V)and andthe thesensor sensor output the King’s law, the relationship between theairflow airflow (denoted asasV) V) output voltage 2 n 2 n (denoted asas U) can be formulated as bV ,a, b, constants and n are constants that were (denoted as U) can formulated as where a,where b,and andna,nare are constants that were determined (denoted U) canbebe formulated asU U2 ==Uaa ++=bV bVan,,+where b, that were determined determined through the least squares estimation. through the least squares estimation. through the least squares estimation.

Figure 8. Experimental setup for airflow sensor calibration. Figure 8. Experimental setup for airflow sensor calibration. Figure 8. Experimental setup for airflow sensor calibration.

7 31744 7

Sensors 2015, 15, 31738–31750 Sensors 2015, 15, page–page Sensors 2015, 15, page–page 1.6 1.6

1.110 1.110

1.5 1.5

1.105 1.105

U(V) U(V)

1.3 1.3

U(V) U(V)

U UFit Fit

1.4 1.4

U=(1.20+0.385V0.52)0.5 U=(1.20+0.385V0.52)0.5

1.100 1.100 1.095 1.095

1.2 1.2

1.090 1.090

1.1 1.1 0 0

2 2

4 6 Volume 4 Flow Rate6 (L/min) Volume Flow(a)Rate (L/min) (a)

8 8

10 10

1.085 1.085 0 0

10 10

U(V) U(V) 20 30 20 Time(min) 30 (b) Time(min) (b)

40 40

50 50

60 60

Figure 9. 9. (a) Calibration of of airflow airflow sensor; sensor; and and (b) (b) stability stability of of the the sensor. sensor. Figure (a) Calibration Calibration Figure 9. (a) of airflow sensor; and (b) stability of the sensor.

The dynamic response of the hot-film sensor was tested to verify dynamically monitoring the The dynamic response of the hot-film sensor was tested to verify dynamically monitoring the respiratory airflow. An electrical excitation method was applied to measure the response time of the respiratory airflow. An electrical excitation method was applied to measure the response time of the sensor. A schematic view of the conditioning circuit is shown in Figure 10a. A source of 10 mV square sensor. A schematic view of the conditioning circuit is shown in Figure 10a. A source source of 10 mV square square pulse voltage in serials with a limiting resistor (10 KΩ) generated a step excitation to the sensor. The pulse voltage resistor (10(10 KΩ) generated a step excitation to the The voltage in inserials serialswith witha alimiting limiting resistor KΩ) generated a step excitation to sensor. the sensor. dynamic response of the sensor was tested and shown in Figure 10b. The experimental result dynamic response of the sensor waswas tested and The dynamic response of the sensor tested andshown shownininFigure Figure10b. 10b. The The experimental experimental result indicates the response time of the sensor is about 140 ms, which is consistent with the bandwidth of indicates the which is is consistent with thethe bandwidth of the response response time timeof ofthe thesensor sensorisisabout about140 140ms, ms, which consistent with bandwidth 7.2 Hz. A low-pass filter combined with an amplifier provided a 3 dB bandwidth of about 7.2 Hz. 7.2 Hz.Hz. A low-pass filter filter combined with an amplifier provided a 3 dB abandwidth of aboutof7.2 Hz. of 7.2 A low-pass combined with an amplifier provided 3 dB bandwidth about This bandwidth can fulfill the typical respiratory rate of an adult which is 12–20 bpm [30]. The sample This bandwidth can fulfillcan the typical respiratory rate of an adult which is 12–20 bpm The sample 7.2 Hz. This bandwidth fulfill the typical respiratory rate of an adult which is [30]. 12–20 bpm [30]. rate of the monitoring system was set as 100 Hz. More experiments demonstrated that the flow range rate the monitoring was set as 100was Hz. set More demonstrated the flow range The ofsample rate of thesystem monitoring system as experiments 100 Hz. More experiments that demonstrated that of the developed hot-film sensor covered 0–60 L/min, the temperature drift was less than ±3% in the of developed sensor hot-film covered sensor 0–60 L/min, the0–60 temperature drift was less than ±3% in less the thethe flow range of hot-film the developed covered L/min, the temperature drift was ambient temperature variation from 25 °C to 50 °C [24]. The benefits˝ of using this integrated micro ˝ ambient temperature variation from 25 °C to 50 °Cfrom [24].25 TheCbenefits of [24]. using The this benefits integrated than ˘3% in the ambient temperature variation to 50 C of micro using hot-film flow sensor lie in: flexibility, easy fabrication and packaging, high sensitivity, capability to hot-film flow sensor in: flexibility, easy fabrication and packaging, high sensitivity, capability to this integrated microliehot-film flow sensor lie in: flexibility, easy fabrication and packaging, high detect comprehensive respiratory parameters, and being extremely sensitive to weak flow. The detect comprehensive parameters,respiratory and beingparameters, extremely and sensitive weak flow. The sensitivity, capability torespiratory detect comprehensive being to extremely sensitive flexible structure makes the sensor capable to be attached on curved surfaces. flexible thestructure sensor capable to be attached on curved surfaces.on curved surfaces. to weakstructure flow. Themakes flexible makes the sensor capable to be attached

Rb Rb

Ra Ra

+ +-

10K 10K

Oscilloscope Oscilloscope

Rt Rt Rh Rh

Rc Rc

(a) (a)

Figure 10. (a) Schematic conditioning circuit of sensor system; and (b) tested dynamic response. Figure Figure 10. 10. (a) (a) Schematic Schematic conditioning conditioning circuit circuit of of sensor sensor system; system; and and (b) (b) tested tested dynamic dynamic response. response.

3.3. Monitoring of Respiration 3.3. 3.3. Monitoring Monitoring of of Respiration Respiration An experiment of respiration simulation was conducted. Figure 11 shows the experimental setup An of simulation conducted. Figure 11 shows the experimental setup An experiment experiment of respiration respiration simulation was was conducted. Figure thecontroller experimental setup for respiration simulation. An air compressor (FI-Tech AT80/38) and11 anshows airflow (Sevenstar for respiration simulation. An air compressor (FI-Tech AT80/38) and an airflow controller (Sevenstar for respiration An air compressor (FI-Tech AT80/38)rate andof an12 airflow (Sevenstar CS230A) were simulation. used as a respiration generator. A respiratory bpm controller with a peak-to-peak CS230A) were used as aa respiration generator. A respiratory rate of 12 bpm with aa peak-to-peak CS230A) were used as respiration generator. A respiratory rate of 12 bpm with peak-to-peak airflow from 3.5 L/min to 13.9 L/min was generated and blew over the upper lip of the face model. airflow from 3.5 L/min L/min to L/min was generated and blew the lip face model. airflow fromwas 3.5 to 13.9 13.9 wasby generated blewover over the upper upper lip of of the the model. The airflow controlled andL/min recorded a PC, andand the airflow signals detected by the face SSRM were The airflow was controlled and recorded by a PC, and the airflow signals detected by the SSRM were The airflow was controlled and recorded by a PC, and the airflow signals detected by the SSRM were transmitted wirelessly to the PC as well. Figure 12 demonstrates a comparison of tested airflow and transmitted wirelessly to PC as Figure demonstrates aa comparison of airflow and transmitted to the the as well. well. Figureby12 12the demonstrates comparison of tested tested airflow and actual flow. wirelessly It is seen that thePC airflow detected SSRM agreed very well with the actual airflow. actual flow. It is seen that the airflow detected by the SSRM agreed very well with the actual airflow. actual is respiratory seen that the airflow detected by the SSRM agreed wellthe with the actual airflow. Table 2flow. lists It the parameters calculated from tested data very by using algorithm described Table 2 lists the respiratory parameters calculated from tested data by using the algorithm described in Section 2. in Section 2. 31745 8 8

Sensors 2015, 15, 31738–31750

Table 2015, 2 lists respiratory Sensors 15,the page–page Sensors 2015, 15, page–page

parameters calculated from tested data by using the algorithm described

in Section 2.

Figure Experimental Figure 11. 11. Experimentalsetup setupfor forrespiration respiration simulation. simulation. Figure 11. Experimental setup for respiration simulation. 16 16

Volume Flow Rate(L/min) Volume Flow Rate(L/min)

Actual Actual

Tested Tested

14 14 12 12 10 10 8 8 6 6 4 4 2 2 0 0

20 20

40 40

60 60

Time(s) Time(s)

80 80

100 100

120 120

Figure 12. 12. Simulation Simulationrespiration respiration signal. signal. Figure Figure 12. Simulation respiration signal. Table 2. 2. Test Test results results of of respiratory respiratory parameters. parameters. Table Table 2. Test results of respiratory parameters.

Parameters Parameters Parameters

Tested value

Tested Testedvalue value Actualvalue value Actual Actual value

MV (L) MV MV(L) (L) 8.0 8.0 8.0 8.3 8.3 8.3

TVTV (L) TV (L)(L) 0.670.67 0.67 0.690.69 0.69

PIFPIF (L/min) (L/min) PIF (L/min) 14.214.2 14.2 13.913.9 13.9

−1) RR −1)) RR(min (min´1 RR (min 12 12 12 12 12 12

To demonstrate diagnosing respiratory diseases, a sleep apnea recognition experiment was To demonstrate demonstrate diagnosing diagnosing respiratory respiratory diseases, diseases, a sleep apnea recognition experiment was To conducted. The experiments on human subjects were approved by Ethical Review Board of Tsinghua conducted. The experiments on human subjects were approved by Ethical Review Board of Tsinghua University (No. 20140084). The experiment for recognizing apnea and hypopnea was conducted in recognizing apnea and hypopnea was conducted in University (No. 20140084). 20140084).The Theexperiment experimentforfor recognizing apnea and hypopnea was conducted the way that at room temperature the subject wore the SSRM and breathed with simulated apneas thethe wayway thatthat at room temperature the subject wore wore the SSRM and breathed with simulated apneas in at room temperature the subject the SSRM and breathed with simulated and hypopneas while the respiratory flow was recorded. The proposed algorithm was applied to and hypopneas while the respiratory flow wasflow recorded. The proposed applied to apneas and hypopneas while the respiratory was recorded. The algorithm proposed was algorithm was recognize apnea and hypopnea from the respiratory flow data detected in real-time. Experimental recognizetoapnea and hypopnea the respiratory data detected in real-time. applied recognize apnea andfrom hypopnea from theflow respiratory flow data detected Experimental in real-time. results are shown in Figure 13, where the apnea and hypopnea were recognized and the numbers of results are shown in Figure 13, where the13, apnea andthe hypopnea were recognized the numbers of Experimental results are shown in Figure where apnea and hypopnea wereand recognized and the apnea and hypopnea episodes (AHI) were counted automatically. AHI is an index used to indicate apnea and hypopnea episodes (AHI) were counted automatically. AHI is an index used to indicate numbers of apnea and hypopnea episodes (AHI) were counted automatically. AHI is an index used the severity of sleep apnea, which is represented by the number of apnea and hypopnea occurring the severity of sleep apnea, which is represented by the number of apnea and hypopnea occurring per hour in sleep (normal 0–5, mild 5–15, moderate 15–30, severe >30). per hour in sleep (normal 0–5, mild 5–15, moderate 3174615–30, severe >30).

9

Sensors 2015, 15, 31738–31750

to indicate the severity of sleep apnea, which is represented by the number of apnea and hypopnea occurring in sleep (normal 0–5, mild 5–15, moderate 15–30, severe >30). Sensorsper 2015,hour 15, page–page

Sensors 2015, 15, page–page

Figure 13. Apnea and hypopnea in real-time respiration signal.

Figure 13. Apnea and hypopnea in real-time respiration signal.

3.4. Monitoring of Oxygen Saturation and Respiration

3.4. Monitoring of Oxygen Saturation and Respiration

Portable Monitoring Task Force of American Academy of Sleep Medicine (AASM) has made a

Figure Apnea hypopnea in real-time respiration recommendation: a portable OSA monitor uses oximetry as one of the signal. monitoring channels has [10,31]. Portable Monitoring Task 13. Force ofand American Academy of Sleep Medicine (AASM) made a When an OSAaevent occurs, themonitor patient cannot inhale oxygen period of time, therefore recommendation: portable OSA uses oximetry as onefor ofathe monitoring channels the [10,31]. 3.4. Monitoring of Oxygen Saturation and Respiration arterial oxygen saturation (SpO 2) will gradually reduce. This is a distinct sign for identifying the When an OSA event occurs, the patient cannot inhale oxygen for a period of time, therefore the apnea or hypopnea event. Combination of oximetry and SSRM to simultaneously monitor SpO2 aand Portable Monitoring Task Force of American Academy of Sleep Medicine (AASM) has made arterial oxygen saturation (SpO 2 ) will gradually reduce. This is a distinct sign for identifying the respiration enables amore precise diagnosis An experiment conducted by using a wrist recommendation: portable OSA monitor for usesOSAS. oximetry as one of thewas monitoring channels [10,31]. apnea or hypopnea event. Combination of oximetry and SSRM to simultaneously monitor SpO2 When an OSA(CONTEC, event occurs, the patient inhale oxygen time, therefore the pulse oximeter CMS50F) worn cannot on the subject’s fingerfor anda aperiod SSRM of attached on their upper and respiration enables more precise diagnosis for OSAS. Anbreath was conducted by arterial oxygen saturation (SpO 2) will gradually This isexperiment a distinct sign forapnea identifying theusing a lip. During the experiment, the subject artificiallyreduce. held their to simulate three times. wrist Each pulse oximeter (CONTEC, CMS50F) the finger and SSRM attached apnea or hypopnea event. ofworn oximetry andsubject’s SSRM to 14 simultaneously monitor SpO2 and time the breath was Combination held for about one on minute. Figure shows thea tested results of on the their upperdesaturation lip. During the experiment, the2,subject artificially held their to simulate respiration enables more precise diagnosis for OSAS. experiment wasbreath conducted by using a apnea wrist three and resaturation of SpO as well as theAn inhale and exhale waveform of the respiration. pulse oximeter CMS50F) on the subject’s finger and a 14 SSRM attached their upper times.The Each time the(CONTEC, breath was about one minute. Figure shows the on tested results results indicate the index ofheld SpOfor 2worn went down after an apnea occurring and exhibited a time delay.of the lip. During experiment, subject artificially held their and breath to simulate apnea of three times. These experimental results were repeatable, and the correction between the incidences of apnea and desaturation and the resaturation ofthe SpO , as well as the inhale exhale waveform the respiration. 2 Each time the breath was held forAs about one minute. Figure 14 shows the tested results ofinthe SpO 2 decreasing is higher than 95%. mentioned above, no-breath or low-breath will result SpO The results indicate the index of SpO2 went down after an apnea occurring and exhibited a2 time desaturation and resaturation of SpO2, as well as the inhale and exhale waveform of the respiration. Normal index results of SpO2 were rangesrepeatable, from 95% to and 100%.the It was reportedbetween that oximetry analyzed of delay.decreased. These experimental correction the incidences The results indicate the index of SpO 2 went down after an apnea occurring and exhibited a time delay. by counting the number of arterial oxygen desaturations index for 4% concluded the OSA, but the apnea and is were higher than 95%. Ascorrection mentioned above, orapnea low-breath will 2 decreasing TheseSpO experimental results repeatable, and the between theno-breath incidences of and diagnosis sensitivity was 40% and specificity was 98%. Serious OSAS patients suffer frequent result in SpO Normal index of SpO2above, ranges from 95% to 100%.willItresult was in reported that SpO 2 decreasing is higher than 95%. As mentioned no-breath or low-breath SpO2 2 decreased. occurrences of apneas accompanied by oxygen desaturation during sleep. A combination of decreased. Normal index of SpO ranges from 95% to 100%. It was reported that index oximetry oximetry analyzed by counting the2number of arterial oxygen desaturations foranalyzed 4% concluded respiratory and SpO2 monitoring could improve the accuracy of diagnosis for OSAS. by counting the number of arterial oxygen desaturations index for 4% concluded the OSA, but the suffer the OSA, but the diagnosis sensitivity was 40% and specificity was 98%. Serious OSAS patients diagnosis sensitivity was 40% and specificity was 98%. Serious OSAS patients suffer frequent frequent occurrences of apneas accompanied by oxygen desaturation during sleep. A combination of occurrences of apneas accompanied by oxygen desaturation during sleep. A combination of respiratory and SpO monitoring could improve the accuracy of diagnosis for OSAS. respiratory and 2SpO 2 monitoring could improve the accuracy of diagnosis for OSAS.

Figure 14. Monitoring of SpO2 and respiration.

Figure 14. Monitoring of SpO2 and respiration.

Figure 14. Monitoring of SpO2 and respiration.

10

31747 10

Sensors 2015, 15, 31738–31750 Sensors 2015, 15, page–page

3.5. 3.5. Monitoring of Sleep Sleep Posture, Posture, Oxygen Oxygen Saturation, Saturation, and and Respiration Respiration Many studies have have demonstrated demonstratedthat thatthe thesleep sleepposture posture has a significant influence Many medical studies has a significant influence on on apneic events, and lateraldecubitus decubituscan canreduce reducethe the frequency frequency of of obstructive respiratory the the apneic events, and lateral respiratory events [31]. [31]. The OSA patients patients at at supine supine posture posture encounter encounter with withapnea/hypopnea apnea/hypopnea events more than than events twice than than those those at at lateral lateral posture. posture. The effect of gravity on the upper airway (UA) of of the the patient patient at at twice the supine posture forfor thethe anatomical andand physiological changes [32].[32]. We the posture isisthe themost mostdominant dominantfactor factor anatomical physiological changes conducted an an experiment to to demonstrate human We conducted experiment demonstratethe therelationship relationshipbetween between body body posture posture and human respiration. The SSRM, to identify identify the the respiration. SSRM, a homemade body posture sensor (using tri-axis accelerometer to posture [13]) and aa CMS50F and monitored continuously in posture CMS50F pulse pulseoximetry oximetrywere werewore woreon onthe thesubject subject and monitored continuously sleep. Figure 15 shows the tested results of body posture, SpO2 and The body posture in sleep. Figure 15 shows the tested results of body posture, SpO2respiration. and respiration. The body sensor was placed front ofinthe chest Z-axis perpendicular to the body. to When posture sensor wasinplaced front of and the the chest and of thesensor Z-axiswas of sensor was perpendicular the the subject supine posture, the posture, magnitude the Z-axisof acceleration was around was 1 g (namely body. Whenwas the at subject was at supine theof magnitude the Z-axis acceleration around 2 ), lateral 2), otherwise posture. Fromposture. the results, wethe canresults, see thewe respiration exhibited non19.8g m/s (namely 9.8 m/sat otherwise at lateral From can see the respiration uniformitynon-uniformity at the supine posture. Several apneas Several occurredapneas in the occurred supine process. In contrast, the exhibited at the supine posture. in the supine process. respiration therespiration lateral posture good uniformity. No apnea/hypopnea occurred in In contrast,atthe at theexhibited lateral posture exhibited good uniformity. No event apnea/hypopnea this process. experimental demonstrated that the lateral posture prevent event occurred The in this process. Theresults experimental results demonstrated that the lateralcan posture can apnea/hypopnea happening. prevent apnea/hypopnea happening.

Figure 15. 15. Monitoring Monitoring of of sleep sleep posture, posture, SpO SpO2,, and and respiration. respiration. Figure 2

4. Conclusions 4. Conclusions A wireless, non-invasive sensing system for monitoring real-time respiratory flow using a A wireless, non-invasive sensing system for monitoring real-time respiratory flow using a monolithically-integrated flexible hot-film flow sensor incorporated with a smart phone or PC is monolithically-integrated flexible hot-film flow sensor incorporated with a smart phone or PC is presented in the paper. The monitoring unit, like a smart strip, is flexible that can be attached on the presented in the paper. The monitoring unit, like a smart strip, is flexible that can be attached human upper lip. The smart strip is comprised of flexible integrated hot-film flow sensor, on the human upper lip. The smart strip is comprised of flexible integrated hot-film flow sensor, conditioning circuit, rechargeable battery, and Bluetooth4.0 LE module, all of which are packaged in conditioning circuit, rechargeable battery, and Bluetooth4.0 LE module, all of which are packaged in a flexible silicon case. The integrated hot-film flow sensor combines four resistors of a Wheatstone a flexible silicon case. The integrated hot-film flow sensor combines four resistors of a Wheatstone bridge operated in CTD mode on one chip, including a hot-film resistor, a temperature compensating bridge operated in CTD mode on one chip, including a hot-film resistor, a temperature compensating resistor and two balancing resistors. As a wearable device, the tube-free configuration makes the resistor and two balancing resistors. As a wearable device, the tube-free configuration makes respiratory monitoring non-invasive and minimal resistance and, therefore, the users could take a the respiratory monitoring non-invasive and minimal resistance and, therefore, the users could more comfortable experience. Low power consumption design is conducted via software and take a more comfortable experience. Low power consumption design is conducted via software hardware optimization to meet the requirements for long-time monitoring, covering at least a sleep and hardware optimization to meet the requirements for long-time monitoring, covering at least a cycle of eight hours. This non-invasive respiratory monitor makes comprehensive respiration sleep cycle of eight hours. This non-invasive respiratory monitor makes comprehensive respiration parameters acquirable, which are important in diagnosis of OSA, COPD, and asthma. In addition of 31748 11

Sensors 2015, 15, 31738–31750

parameters acquirable, which are important in diagnosis of OSA, COPD, and asthma. In addition of respiration monitoring, the proposed devices can be also applied in the monitoring of newborns during nutritive suction [33]. Acknowledgments: This work was financially supported by the National High-tech Program “863” of China under the Grant No. 2012AA02A604. Author Contributions: Peng Jiang developed the system, conducted experiments, and wrote the article under the supervision of Rong Zhu. Shuai Zhao designed and fabricated the monolithically integrated flexible hot-film flow sensor. Rong Zhu proposed the original idea. Conflicts of Interest: The authors declare no conflict of interest.

References 1.

2. 3. 4. 5.

6.

7.

8.

9.

10.

11. 12. 13.

14. 15.

World Health Organization. The 10 Leading Causes of Death in the World, 2000 and 2012; World Health Organization: Geneva, Switzerland, 2014. Available online: http://www.who.int/mediacentre/ factsheets/fs310/en/ (accessed on 15 May 2014). Jain, C.D.; Bhaskar, D.J.; Kalra, M.; Gandhi, R.; Bohra, P. Recent Improvements in the Management of Obstructive Sleep Apnea: The Dental Perception. Int. J. Adv. Health Sci. 2015, 1, 7–10. Vaessen, T.J.; Overeem, S.; Sitskoorn, M.M. Cognitive complaints in obstructive sleep apnea. Sleep Med. Rev. 2015, 19, 51–58. [CrossRef] [PubMed] Yaggi, H.K.; Araujo, A.B.; McKinlay, J.B. Sleep duration as a risk factor for the development of type 2 diabetes. Diabetes Care 2006, 29, 657–661. [CrossRef] [PubMed] National Institutes of Health. National Heart, Lung and Blood Institute (NHLBI). Guidelines for the Diagnosis and Management of Asthma (EPR-3); NHLBI: Bethesda, MD, USA. Available online: http://www.nhlbi.nih.gov/guidelines/asthma (accessed on 10 Feburary 2011). Kushida, C.A.; Littner, M.R.; Morgenthaler, T.; Alessi, C.A.; Bailey, D.; Coleman, J., Jr.; Friedman, L.; Hirshkowitz, M.; Kapen, S.; Kramer, M. Practice parameters for the indications for polysomnography and related procedures: An update for 2005. Sleep 2005, 28, 499–521. [PubMed] Oliveira, M.G.; Nery, L.E.; Santos-Silva, R.; Sartori, D.E.; Alonso, F.F.; Togeiro, S.M.; Tufik, S.; Bittencourt, L. Is portable monitoring accurate in the diagnosis of obstructive sleep apnea syndrome in chronic pulmonary obstructive disease? Sleep Med. 2012, 13, 1033–1038. [CrossRef] [PubMed] Chen, H.; Lowe, A.A.; Bai, Y.; Hamilton, P.; Fleetham, J.A.; Almeida, F.R. Evaluation of a portable recording device (apnealink™) for case selection of obstructive sleep apnea. Sleep Breath. 2009, 13, 213–219. [CrossRef] [PubMed] Collop, N.A.; Anderson, W.M.; Boehlecke, B.; Claman, D.; Goldberg, R.; Gottlieb, D.; Hudgel, D.; Sateia, M.; Schwab, R. Clinical guidelines for the use of unattended portable monitors in the diagnosis of obstructive sleep apnea in adult patients. J. Clin. Sleep Med. 2007, 3, 737–747. [PubMed] Petrovi´c, M.; Petrovic, J.; Daniˇci´c, A.; Vukˇcevi´c, M.; Bojovi´c, B.; Hadžievski, L.; Allsop, T.; Lloyd, G.; Webb, D. Non-invasive respiratory monitoring using long-period fiber grating sensors. Biomed. Opt. Express. 2014, 5, 1136–1144. [CrossRef] [PubMed] Cao, Z.; Zhu, R.; Que, R.-Y. A wireless portable system with microsensors for monitoring respiratory diseases. IEEE Trans. Biomed. Eng. 2012, 59, 3110–3116. [PubMed] Folke, M.; Cernerud, L.; Ekström, M.; Hök, B. Critical review of non-invasive respiratory monitoring in medical care. Med. Biol. Eng. Comput. 2003, 41, 377–383. [CrossRef] [PubMed] Dziuda, L. Fiber-optic sensors for monitoring patient physiological parameters: A review of applicable technologies and relevance to use during magnetic resonance imaging procedures. J. Biomed. Opt. 2015, 20, 010901. [CrossRef] [PubMed] Pereira, C.; Yu, X.; Czaplik, M.; Rossaint, R.; Blazek, V.; Leonhardt, S. Remote monitoring of breathing dynamics using infrared thermography. Biomed. Opt. Express 2015, 6, 4378–4394. [CrossRef] [PubMed] Schena, E.; Massaroni, C.; Saccomandi, P.; Cecchini, S. Flow measurement in mechanical ventilation: A review. Med. Eng. Phys. 2015, 37, 257–264. [CrossRef] [PubMed]

31749

Sensors 2015, 15, 31738–31750

16.

17. 18. 19. 20. 21. 22.

23. 24. 25. 26. 27.

28. 29.

30. 31.

32.

33.

Snepvangers, Y.; de Winter, P.; Burger, H.; Brouwers, H.A.; Bogaard, J.M.; van der Ent, K. Correction factors for oxygen and gas flow-rate effects on neonatal Fleisch and Lilly pneumotachographs. Pediat. Crit. Care Med. 2003, 4, 227–232. [CrossRef] [PubMed] Tardi, G.; Massaroni, C.; Saccomandi, P.; Schena, E. Experimental Assessment of a Variable Orifice Flowmeter for Respiratory Monitoring. J. Sens. 2015. [CrossRef] Kuo, J.T.; Yu, L.; Meng, E. Micromachined thermal flow sensors—A review. Micromachines 2012, 3, 550–573. [CrossRef] Nguyen, N. Micromachined flow sensors—A review. Flow Meas. Instrum. 1997, 8, 7–16. [CrossRef] Sturm, H.; Lang, W. Membrane-based thermal flow sensors on flexible substrates. Sens. Actuators A Phys. 2013, 195, 113–122. [CrossRef] Petropoulos, A.; Pagonis, D.N.; Kaltsas, G. Flexible PCB-MEMS flow sensor. Procedia Eng. 2012, 47, 236–239. [CrossRef] Sturm, H.; Brauns, E.; Froehner, K.; Lang, W.; Buchner, R. Thermoelectric Flow Sensors on Flexible Substrates and Their Integration Proces. In Proceedings of the IEEE Sensors, Kona, HI, USA, 1–4 November 2010; pp. 575–579. Bowman, B.; Stasz, P. Method of Making a Respiration Sensor. U.S. Patent 5,832,592, 10 November 1998. Zhao, S.; Zhu, R. Monolithic integration of a flexible micro thermal flow sensor. In Applied Mechanics and Materials; Trans Tech Publication: Beijing, China, 2015; pp. 89–92. Bruun, H.H. Hot-Wire Anemometry Principles and Signal Analysis. Meas. Sci. Technol. 1995, 7. [CrossRef] Gomez, C.; Oller, J.; Paradells, J. Overview and evaluation of bluetooth low energy: An emerging low-power wireless technology. Sensors 2012, 12, 11734–11753. [CrossRef] Devita, G.; Wong, A.; Dawkins, M.; Glaros, K.; Kiani, U.; Lauria, F.; Madaka, V.; Omeni, O.; Schiff, J.; Vasudevan, A. A 5 mw Multi-Standard Bluetooth Le/Ieee 802.15. 6 Soc for Wban Applications. In Proceedings of the 40th European Solid State Circuits Conference (ESSCIRC), Venice, Italy, 22–26 September 2014; IEEE: New York, NY, USA, 2014; pp. 283–286. Zuurbier, M.; Hoek, G.; van den Hazel, P.; Brunekreef, B. Minute ventilation of cyclists, car and bus passengers: An experimental study. Environ. Health 2009, 8, 48. [CrossRef] [PubMed] Harrison, K.L.; Bogard, D.G. Comparison of Rans Turbulence Models for Prediction of Film Cooling Performance. In Proceedings of the ASME Turbo Expo 2008: Power for Land, Sea, and Air, Berlin, Germany, 9–13 June 2008; pp. 1187–1196. Grossman, E.; Grossman, A.; Schein, M.; Zimlichman, R.; Gavish, B. Breathing-control lowers blood pressure. J. Hum. Hypertens. 2001, 15, 263–269. [CrossRef] [PubMed] Collop, N.A.; Tracy, S.L.; Kapur, V.; Mehra, R.; Kuhlmann, D.; Fleishman, S.A.; Ojile, J.M. Obstructive sleep apnea devices for out-of-center (ooc) testing: Technology evaluation. J. Clin. Sleep Med. JCSM 2011, 7, 531–548. [CrossRef] [PubMed] Ozeke, O.; Erturk, O.; Gungor, M.; Hizel, S.B.; Aydin, D.; Celenk, M.K.; Dincer, H.; Ilicin, G.; Ozgen, F.; Ozer, C. Influence of the right-versus left-sided sleeping position on the apnea-hypopnea index in patients with sleep apnea. Sleep Breath. 2012, 16, 617–620. [CrossRef] [PubMed] Tamilia, E.; Taffoni, F.; Formica, D.; Ricci, L.; Schena, E.; Keller, F.; Guglielmelli, E. Technological solutions and main indices for the assessment of newborns' Nutritive Sucking: A review. Sensors 2014, 14, 634–658. [CrossRef] [PubMed] © 2015 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons by Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).

31750