An Electronic System for the Contactless Reading of ECG ... - MDPI

3 downloads 0 Views 4MB Size Report
Oct 28, 2017 - Unit of Electronics for Sensor Systems, Department of Engineering Campus ... Keywords: capacitive sensors; ECG; contactless; biomedical ...
sensors Article

An Electronic System for the Contactless Reading of ECG Signals Francesca Romana Parente 1 ID , Marco Santonico 2 , Alessandro Zompanti 2 , Mario Benassai 3 , Giuseppe Ferri 1 ID , Arnaldo D’Amico 4 and Giorgio Pennazza 2, * 1 2 3 4

*

Department of Industrial and Information Engineering and Economics, University of L’Aquila, 67100 L’Aquila, Italy; [email protected] (F.R.P.); [email protected] (G.F.) Unit of Electronics for Sensor Systems, Department of Engineering Campus Bio-Medico University of Rome, 00128 Rome, Italy; [email protected] (M.S.); [email protected] (A.Z.) ALTEC S.p.A., Aerospace Logistics and Technology Engineering Company, 10146 Torino, Italy; [email protected] Department of Electronic Engineering, University of Rome Tor Vergata, 00133 Rome, Italy; [email protected] Correspondence: [email protected]

Received: 25 September 2017; Accepted: 27 October 2017; Published: 28 October 2017

Abstract: The aim of this work is the development of a contactless capacitive sensory system for the detection of (Electrocardiographic) ECG-like signals. The acquisition approach is based on a capacitive coupling with the patient body performed by electrodes integrated in a front-end circuit. The proposed system is able to detect changes in the electric charge related to the heart activity. Due to the target signal weakness and to the presence of other undesired signals, suitable amplification stages and analogue filters are required. Simulated results allowed us to evaluate the effectiveness of the approach, whereas experimental measurements, recorded without contact to the skin, have validated the practical effectiveness of the proposed architecture. The system operates with a supply voltage of ±9 V with an overall power consumption of about 10 mW. The analogue output of the electronic interface is connected to an ATmega328 microcontroller implementing the A/D conversion and the data acquisition. The collected data can be displayed on any multimedia support for real-time tracking applications. Keywords: capacitive sensors; ECG; contactless; biomedical electronics; wearable

1. Introduction In recent years, there has been a growing interest in the detection of signals from the human body [1–4]. Typically, these measurements are performed by the use of Ag/AgCl electrodes with conductive gels in direct contact with the skin in order to transduce the body surface ion current into an electron current [5–7]. These bioelectric signals represent the result of the synchronous electrochemical activity of the nervous tissue cells. Hence, a signal picked up from the human body consists of an analogue signal, which is characterized by low intensity, low frequency and narrow bandwidth. In particular, traditional electrocardiography, ECG, is able to record those potentials on the body surface, directly reflecting the electrical cardiac activity with an amplitude of few mV with frequencies in the range of 0.01–150 Hz [8]. In addition, in the electrocardiographic acquisition process, it is necessary to filter out some undesirable signals, such as, for example, those caused by the power-line supply (50 Hz in Europe, 60 Hz in the USA) or those coming from other bioelectrical sources [9]. Typically, the use of contact standard electrodes requires a proper preparation of the body surface due to the presence of body hair, abrasions or dirt [10]. Moreover, prolonged use of the gel can cause irritation, as well as the loss of its conductive capacity, caused by a drying out over time [11]. Sensors 2017, 17, 2474; doi:10.3390/s17112474

www.mdpi.com/journal/sensors

Sensors 2017, 17, 2474 Sensors 2017, 17, 2474

2 of 10 2 of 10

can cause irritation, as well as the loss of its conductive capacity, caused by a drying out over time [11]. In order to overcome these drawbacks, dry contact electrodes have been widely investigated for In order to overcome these drawbacks, dry contact electrodes have been widely investigated for long-term heart monitoring applications [12,13]. In some applications, these transducers have shown long-term heart monitoring applications [12,13]. In some applications, these transducers have shown better performance with respect to traditional wet gel electrodes in terms of electrode–skin interface better performance with respect to traditional wet gel electrodes in terms of electrode–skin interface impedance, signal intensity and size [14]. impedance, signal intensity and size [14]. In the last years, non-contact electrodes have been considered as an alternative option with In the last years, non-contact electrodes have been considered as an alternative option with respect respect to contact electrodes [15,16]. In particular, in the literature, capacitive sensors performing a to contact electrodes [15,16]. In particular, in the literature, capacitive sensors performing a coupling coupling with a human body have been explored [17–21]: whilst they still have not replaced with a human body have been explored [17–21]: whilst they still have not replaced traditional wet traditional wet electrodes, they have shown satisfactory results in particular applications, such as in electrodes, they have shown satisfactory results in particular applications, such as in the automotive the automotive environment [20] or in wearable health devices [21]. Having an electrode without environment [20] or in wearable health devices [21]. Having an electrode without direct contact with direct contact with the skin offer several benefits in long-term monitoring applications, such as the skin offer several benefits in long-term monitoring applications, such as stability over time and stability over time and having the capacitive sensor surface electrically insulated [22]. In addition, a having the capacitive sensor surface electrically insulated [22]. In addition, a contactless monitoring contactless monitoring system may be more comfortable for the involved subject, as it is system may be more comfortable for the involved subject, as it is user-friendly, not invasive at all and user-friendly, not invasive at all and it does not require the presence of qualified medical staff. it does not require the presence of qualified medical staff. In this paper, a complete system for the contactless real-time monitoring of ECG signal is In this paper, a complete system for the contactless real-time monitoring of ECG signal is presented. presented. It exploits electrodes performing a capacitive coupling with the patient body that through It exploits electrodes performing a capacitive coupling with the patient body that through a dedicated a dedicated front-end circuit are able to detect electrical charge variations related to the heart front-end circuit are able to detect electrical charge variations related to the heart activity. This work activity. This work represents an optimized version of an already-presented system that shows only represents an optimized version of an already-presented system that shows only simulated results simulated results for a preliminary version, employing one single capacitive electrode [23]. In this for a preliminary version, employing one single capacitive electrode [23]. In this study, two leading study, two leading electrodes and one more as reference (that has to be placed far from the cardiac electrodes and one more as reference (that has to be placed far from the cardiac source) were included source) were included and the front-end was been completely redesigned. and the front-end was been completely redesigned. The promising results obtained here pave the way to wearable solutions for an ECG contactless The promising results obtained here pave the way to wearable solutions for an ECG contactless device via a suitable integration of the system described in the paper. The active filtering effectively device via a suitable integration of the system described in the paper. The active filtering effectively supports the extraction of a clear signal at the frequency of interest, as shown by the experimental supports the extraction of a clear signal at the frequency of interest, as shown by the experimental results reported here, which validate the practical effectiveness of the proposed approach. results reported here, which validate the practical effectiveness of the proposed approach. 2. The The Proposed Proposed System System 2. In this this section, section, the the details details of of the the designed designed system system are are reported. reported. The The block block scheme scheme reported reported in in In Figure 11 summarizes summarizes the the main main stages. stages. The the heart heart activity, activity, through through the the contactless contactless Figure The detection detection of of the electrodes, occurs at the front-end portion (that includes the electrodes and a preamplifier), whose electrodes, occurs at the front-end portion (that includes the electrodes and a preamplifier), whose output signal is then fed to an analogue signal conditioning circuit formed by filters and by an output signal is then fed to an analogue signal conditioning circuit formed by filters and by an amplifier. amplifier. Theoutput amplified signal up by a(µC) microcontroller dealing with the The amplified signaloutput is picked up byisapicked microcontroller dealing with (μC) the analogue-to-digital analogue-to-digital conversion and the digital data storage. conversion and the digital data storage.

Figure 1. 1. The proposed proposed system system block block scheme. scheme. Figure

In order to evaluate the circuit limits, as well as to optimize its performance, simulations of the In order to evaluate the circuit limits, as well as to optimize its performance, simulations of the analogue electronic filter stages were performed by using National Instruments NI Multisim analogue electronic filter stages were performed by using National Instruments NI Multisim software. software. As an operational amplifier, LF411 was employed [24]. All the active elements in the As an operational amplifier, LF411 was employed [24]. All the active elements in the electronic interface electronic interface were supplied at ±9 V. were supplied at ±9 V.

Sensors 2017, 17, 2474 Sensors Sensors2017, 2017,17, 17,2474 2474

3 of 10 33of of10 10

Front-End Block 2.1.2.1. TheThe Front-End Block 2.1. The Front-End Block ECG monitoring through capacitive measurement In this work, ECG monitoring through requires dedicatedelectrode. electrode. this work, ECG monitoring throughcapacitive capacitivemeasurement measurementrequires requiresaaadedicated dedicated electrode. InIn this work, an electrode made up of a single conductive (copper) layer was designed. This electrode, when electrode made a singleconductive conductive(copper) (copper)layer layer was was designed. designed. This when an an electrode made upup of of a single Thiselectrode, electrode, when brought close the subject’s chest (see Figure 2), can be modelled brought close to the subject’schest chest(see (seeFigure Figure2), 2),can canbe be modelled modelled through through aaacapacitor capacitor brought close toto the subject’s through capacitorCCelCelwhose whose elwhose value, according to the parallel plate capacitor theory, depends on the distance, on the size and on value, according to the parallel plate capacitor theory, depends distance, size and on value, according to the parallel plate capacitor theory, depends onon thethe distance, onon thethe size and on the the dielectric material between the two conductors [25]. the dielectric material between theconductors two conductors dielectric material between the two [25]. [25].

Figure Electrode positioning Figure on the human body. Figure2.2.2.Electrode Electrodepositioning positioningon onthe thehuman humanbody. body.

Having considered aasingle layer cotton T-shirt (dielectric constant: = 1.4) with a thickness Having considered single layer cottonT-shirt T-shirt(dielectric (dielectricconstant: constant:εεεcotton cotton = 1.4) with a thickness Having considered a single layer cotton cotton = 1.4) with a thickness 2, an of roughly 0.5 mm and a fixed size for the electrodes of about 65 cm equivalent capacitance 2 of roughly 0.5 mm a fixed for electrodes the electrodes of about 652 , cm , an equivalent capacitance of roughly 0.5 mm andand a fixed size size for the of about 65 cm an equivalent capacitance value value of about 150 pF was estimated. A high valued resistor (few GΩ), connected to ground, value of about 150 pF was estimated. A high valued resistor (few GΩ), connected to ground,in inseries series of about 150 pF was estimated. A high valued resistor (few GΩ), connected to ground, in series with with withthis thiscapacitor, capacitor,forms formsaapassive passivehigh-pass high-passcell cellat atvery verylow lowfrequencies frequencies(of (ofabout about11Hz). Hz). this capacitor, forms aboth passivematerial high-passr) cell at very low frequencies (of about 1 in Hz). AAvariation variationin in boththe the material(ε (εr)and andthe thethickness thickness(t) (t)of ofthe thet-shirt t-shirtresults results inaamodification modificationof of A variation in both the material (ε ) and the thickness (t) of the t-shirt results in parameters a modification of r of the overall transfer function on these two the cutoff frequency. The dependence the cutoff frequency. The dependence of the overall transfer function on these two parameters has has thebeen cutoff frequency. The dependence ÷of1.6, the overall transfer function on these two parameters has been beenverified verifiedby bysimulation—ε simulation—εr r=1.3 =1.3 ÷ 1.6,tt==1,1,0.75, 0.75,0.50, 0.50,0.25 0.25mm—these mm—theseranges rangeshave havebeen beenfound foundin in verified by simulation—ε =1.3 ÷ 1.6, t = 1, 0.75, 0.50, of 0.25 mm—these ranges haveboth beenthe found in the r the literature [18,26,27]. At low frequency, the range interest of the ECG signal, material the literature [18,26,27]. At low frequency, the range of interest of the ECG signal, both the material literature At low frequency, the range ofItItinterest the ECG that, signal, both the material and and thickness the function. isisworth remarking the thickness andthe the[18,26,27]. thicknessinfluence influence thetransfer transfer function. worthof remarking that,although although the thickness theeffect thickness influence the transfer function. It is worth remarking although the thickness effect seems to significant than material, they do prevent effective signal effect seems to be be more more significant than the the material, they both boththat, do not not prevent effective signal seems to be more significant than the material, they both do not prevent effective signal capturing. capturing. capturing. The high-pass filter isis present of leading electrodes, as in 3,3, The high-pass filter is present on on both of the two leading electrodes, as shown in Figure 3, which The high-pass filter present on both both of the the two two leading electrodes, as shown shown in Figure Figure which also includes the first analogue front-end amplifier. alsowhich includes first analogue front-end amplifier. also the includes the first analogue front-end amplifier.

Figure 3.3.The front-end Figure3. Thefront-end front-endblock. block. Figure The block.

Sensors 2017, 17,17, 2474 Sensors 2017, 2474 Sensors 2017, 17, 2474

4 10 of 44of of 10 10

The Thesignal signalassociated associatedwith withcardiac cardiacactivity activitytypically typicallyshows showsvery verylow lowvoltage voltagelevels levelsin inaastandard standard The signal associated with activity typically shows very voltage in a standard contact acquisition and also lower therefore the presence of aa preamplifier isis contact acquisition and isis alsocardiac lower in in this this approach, approach, therefore thelow presence oflevels preamplifier contact acquisition and is lower thisresistor, approach, therefore presenceinstrumentation of a preamplifieramplifier is mandatory. mandatory. In to match the high aahigh input impedance isis mandatory. Inorder order toalso match thein high resistor, high inputthe impedance instrumentation amplifier In included order to (INA128 match the high resistor, a high input impedance instrumentation amplifier [28]). This amplifier shows aahigh common rejection (CMRR). single included (INA128 [28]). This amplifier shows high commonmode mode rejectionratio ratio (CMRR).isA Aincluded single (INA128 [28]). ThisRRG1 amplifier shows chosen achosen high resistance common rejection ratio (CMRR). Again single external external resistor value 12 the amplifier to ≈≈5, external resistor G1,,whose whosedefault default resistancemode valueisis 12kΩ, kΩ,sets sets the amplifier gain toG1 G1 5, resistor R , whose default chosen resistance value is 12 kΩ, sets the amplifier gain to G1 ≈ 5, i.e., about i.e., i.e., about about 14 dB. dB. Due Due to to both both the the interpersonal interpersonal variability variability and and the the measurement measurement conditions, conditions, this this G1 14 might be in cases. we included allows value might be unsuitable unsuitable in some some cases. Hence, Hence, we included aa trimmer trimmer (R (RG1 G1), ), which which allowsbe 14 value dB. Due to both the interpersonal variability and the measurement conditions, this value might adjustment of gain value. In addition, between input terminals, mm,, helps to adjustment of the the gainHence, value.we Inincluded addition,aatrimmer a capacitor capacitor between the input terminals, helpsvalue. to unsuitable in some cases. (RG1 ), whichthe allows adjustment ofCC the gain reduce the magnetic coupling noise. reduce the magnetic coupling noise. In addition, a capacitor between the input terminals, Cm , helps to reduce the magnetic coupling noise. Band-Pass Filter 2.2. The Band-Pass Filter 2.2.2.2. TheThe Band-Pass Filter The first filtering stage active band-pass filter (see Figure which limits The first filtering stage is represented represented by an active band-pass filter (see Figure 4), which limits The first filtering stage is is represented byby anan active band-pass filter (see Figure 4),4), which limits the the bandwidth to the frequencies of interest in the ECG signal [9]. This design choice does not affect the bandwidth to the frequencies of interest in the ECG signal [9]. This design choice does not affect bandwidth to the frequencies of interest in the ECG signal [9]. This design choice does not affect the the helps in of of the acquisition acquisition process (while, indeed, helps in the the highlighting highlighting of the the information information of interest), interest), acquisition processprocess (while,(while, indeed,indeed, it helpsititin the highlighting of the information of interest), since the since the significant the below 30 even ififup itit can be 150 Hz since the most mostpart significant part of of theECG ECG signal lies lies below 30 Hz, evenbe can be up up to[8]. 150Moreover, Hz [8]. [8]. most significant of the part ECG signal lies signal below 30 Hz, even ifHz, it can to 150 Hzto Moreover, its cut-off at low frequencies (around 0.2 Hz) reduces the baseline drift due to respiration. Moreover, its cut-off at low frequencies (around 0.2 Hz) reduces the baseline drift due to respiration. its cut-off at low frequencies (around 0.2 Hz) reduces the baseline drift due to respiration. Furthermore, Furthermore, the the control of the resistors resistors in in the the loop loop circuit circuit allow the gain of the the signal signal in in the the bandwidth, bandwidth, theFurthermore, resistors in the loop circuit allow the gain allow control ofgain the control signal in the bandwidth, which has an which whichhas hasan anamplification amplificationfactor factorof ofabout about55dB, dB,as asshown shownin inFigure Figure5. 5. amplification factor of about 5 dB, as shown in Figure 5.

Figure Active band-pass Figure4.4.4.Active Activeband-pass band-passfilter filterschematic schematiccircuit. circuit. Figure filter schematic circuit.

Figure Active band-pass Figure5.5.5.Active Activeband-pass band-passfilter filtertransfer transferfunction. function. Figure filter transfer function.

Notch Filter 2.3. The Notch Filter 2.3.2.3. TheThe Notch Filter A A second-order second-order Twin-T Twin-T notch notch filter filter centred centred on on the the frequency frequency of of 50 50 Hz Hz obtains obtains the the reduction reduction of of A second-order Twin-T notch filter centred on the frequency of 50 Hz obtains the reduction of the the the power power supply supply interference. interference. In In particular, particular, in in order order to to increase increase the the attenuation attenuation feature, feature, the the power supply interference. In particular, in order to increase the attenuation feature, the standard Twin-T configuration was modified by the addition of extra passive components, Cn and Rn (see Figure 6).

Sensors 2017, 17, 2474 Sensors 2017, 17, 2474

5 of 10 5 of 10

standard Twin-T configuration was modified by the addition of extra passive components, Cn and Rn

standard Sensors 2017,Twin-T 17,6). 2474 configuration was modified by the addition of extra passive components, Cn and 5 of R 10n (see Figure (see Figure 6).

Figure 6. Modified Twin-T notch filter schematic circuit. Twin-T notch notch filter filter schematic schematic circuit. circuit. Figure 6. Modified Twin-T

The values of these two additional components can be set according to (1) and (2): The values of of these these two two additional additional components components can can be be set set according accordingto to(1) (1)and and(2): (2): The values Cn = nC (1) Cn = nC (1) nC (1) RCnn==(R)/2n (2) Rn = (R)/2n (2) where n is a positive integer. In this novelRarchitecture, the cut-off frequency is still given by the (2) n = (R)/2n the cut-off frequency is still given by the where n is a in positive integer. In network this novel(i.e., architecture, components the main passive it is given by fcutoff = 1/2πRC). However, as shown in components the main network itattenuation. is given the byOn fcut-off cutoff = 1/2πRC). However, as shown inb where a in positive integer. In the this novel(i.e., architecture, frequency isfor still given Figure n7,isthe higher the npassive value, higher the the other hand, equal Ra by andthe R Figure 7, the higher the n value, the higher the attenuation. On the other hand, for equal R a and R components the main networkQ(i.e., it is given by fcutoff in = 1/2πRC). However, shown inb values (that in usually setpassive the standard factor), the increase the reduction of theas undesired values (that usually set the standard Q factor), the increase in the reduction of the undesired Figure 7, the higher n value, the higher thein attenuation. On the other hand,tofor equal and Rof b frequency results inthe a lower selectivity, thus a wider stop-band. In order limit theRalosses frequency results in a lower selectivity, thus in a wider stop-band. In order to limit the losses of values (that usually set the standard Q factor), the increase in the reduction of the undesired frequency important components of the target signal, we have chosen n = 1 for the notch filter included in important of thethus target have chosen n = 1 to forlimit the the notch filterofincluded in results in acomponents lower selectivity, in signal, a widerwe stop-band. In order losses important this study. this study. components of the target signal, we have chosen n = 1 for the notch filter included in this study.

Figure 7. Modified notch filter transfer function and its dependence on n. Figure 7. Modified notch filter transfer function and its dependence on n. Figure 7. Modified notch filter transfer function and its dependence on n.

Sensors 2017, 17, 2474 Sensors 2017, 17, 2474

6 of 10 6 of 10

2.4. The Second Amplification Stage 2.4. In Thethe Second Amplification Stage second amplification stage, a higher gain is provided in order to ensure a suitable output signal. For the considered application, an instrumentation amplifier (INA126), with a gain value of In the second amplification stage, a higher gain is provided in order to ensure a suitable output around G2 = 20 (26 dB), was selected. As for the first stage, a trimmer was included for gain signal. For the considered application, an instrumentation amplifier (INA126), with a gain value of adjusting. around G2 = 20 (26 dB), was selected. As for the first stage, a trimmer was included for gain adjusting.

2.5. 2.5.The TheWhole WholeInterface InterfaceSchematic Schematic The Theschematic schematicofofthe theproposed proposedsystem, system,including includingall allthe theanalogue analoguestages, stages,isisreported reportedininFigure Figure8.8. The values of ofpassive passivecomponents components listed in Table A coupling capacitor c was The employed employed values areare listed in Table 1. A 1. coupling capacitor Cc wasCplaced placed at the end of the analogue circuit, to reject the DC components and to restore the baseline of at the end of the analogue circuit, to reject the DC components and to restore the baseline of the the output signal. output signal.

Figure8.8.Circuit Circuitschematic schematicofofthe thewhole wholeanalogue analoguesystem. system. Figure Table 1.Component Componentvalues. values. Table1. Value NameName Value [Ω][Ω] Rin

1G

Name Name R4

Value [Ω] Value [Ω] 8.2 k

Name Name Cm

Value [F] [F] Value 180 p

Rin RG1 1 G 12 k R4 R 8.2 kk 31.82 C1Cm 220180 n p RG1 R1 12 k3.3 M R Ra 31.82 3.3 kk C2C1 100220 n n 1 kk C C2 100100 n n R1 R2 3.3 M33 k Ra Rb 3.3 R3 4.7 k RG2 1.8 k Cc 15 n R2 33 k Rb 1k C 100 n R3 4.7 k RG2 1.8 k Cc 15 n The whole system fabricated was composed by a box containing the electronic interface described and The the capacitive electrodes. All these in Figure 9.the electronic interface whole system fabricated wascomponents composed are by reported a box containing described and the capacitive electrodes. All these components are reported in Figure 9.

Sensors 2017, 17, 2474

7 of 10

Sensors 2017, 17, 2474

7 of 10

Figure 9. System overview (the box containing the electronic board, the electronic board, the three Figure 9. System overview (the box containing the electronic board, the electronic board, the three electrodes connected to the system). electrodes connected to the system).

2.6. Data Acquisition 2.6. Data Acquisition A μC Atmega 328 was used for the output signal analogue-to-digital conversion. Dedicated A µC Atmega 328 was used for the output signal analogue-to-digital conversion. Dedicated software was implemented for data storage as well as for real-time display of the ECG signal software was implemented for data storage as well as for real-time display of the ECG signal acquired acquired μC. In addition, digital filtering handling and further data processing were included so as µC. In addition, digital filtering handling and further data processing were included so as to obtain to obtain a sharp signal by the action of the following digital filter: An = An − En, where En = (αAn−1) + (1 a sharp signal by the action of the following digital filter: An = An − En , where En = (αAn −1 ) + (1 − − α)En−1, An is the n-th sample, En is the n-th sample, and α is the parameter defining the filter α)En− 1 , An is the n-th sample, En is the n-th sample, and α is the parameter defining the filter intensity. intensity. 3. Results 3. Results 3.1. Simulated Results 3.1. Simulated Results Simulations of the whole system for a typical cardiac signal were carried out by NI Multisim. of the whole system a typical cardiac signal were carried out NI Multisim. Two Simulations different source generators were for included in the furtherance of emulating theby input signal at Two different source generators were included in the furtherance of emulating the input signal at each leading electrode; a sinusoidal voltage generator at 50 Hz represents the power-line interference each leading electrode; a sinusoidal voltage generator at 50 Hz represents the power-line while the cardiac waveform was reproduced using a piecewise linear voltage generator. In particular, interference the cardiac waveform reproduced linear voltage generator. the waveformwhile in Figure 10a represents thewas input signal (as ausing sum aofpiecewise the above-mentioned waveforms), In particular, the waveform in Figure 10a represents the input signal (as a sum of the whereas that one in Figure 10b is the output signal. The system provides a considerable reduction of above-mentioned waveforms), whereas that one in Figure 10b is the output signal. The system power-line interference effects. provides a considerable reduction of power-line interference effects. 3.2. Experimental Results 3.2. Experimental Results Experimental measurements were conducted on 10 subjects wearing a single layer cotton T-shirt. Experimental measurements were conducted on using 10 subjects wearingsystem a single cotton Figure 11 reports a typical ECG signal recording acquired the developed withlayer a sampling T-shirt. Figure 11 reports a typical ECG signal recording acquired using the developed system with a rate of 20 ms. For this acquisition, two electrodes were placed on the t-shirt of the control individual. sampling ratesoftware of 20 ms.(SW) For this acquisition, electrodes were placed the t-shirt the control A dedicated interfaced via atwo dedicated Graphical User on Guide (GUI),of displays the individual. A dedicated software (SW) interfaced via aQ-R-S-T dedicated Graphical User Guide ECG-like signal in real time. The wave T and the complex can be easily identified, while(GUI), the P displays thealways ECG-like signal real time. Theit wave T to and the the complex Q-R-S-T can be easily wave is not evident. TheinSW GUI makes possible adjust acquisition parameters while identified, data, whileand the itPalso wave is not evident. SW GUIelaborations makes it possible to adjust the registering allows thealways physician some The preliminary on the signal. acquisition parameters while registering data, and it also allows the physician some preliminary elaborations on the signal.

Sensors 2017, 17, 2474 Sensors 2017, 2017,17, 17,2474 2474 Sensors

8 of 10 of 10 10 88 of

Figure 10. 10. Simulation Simulation of of the (a) (a) input signal signal (b) output output signal. signal. Figure Figure 10. Simulation of the the (a)input input signal(b) (b) output signal.

Figure 11. Heart signal recordings (two electrodes placed on the t-shirt of one of the ten control Figure 11. 11. Heart Heart signal signal recordings recordings (two (two electrodes electrodes placed placed on on the the t-shirt t-shirt of of one one of of the the ten ten control control Figure individuals tested). The ECG-like signal is displayed in real time on the window of the dedicated SW. individuals tested). tested). The TheECG-like ECG-like signal signal is isdisplayed displayedin in real realtime time on on the the window window of of the thededicated dedicatedSW. SW. individuals The detail in the zoom put in evidence that the wave T and the complex QRST can be easily identified, The detail detail in in the the zoom zoom put put in in evidence evidence that that the the wave wave TT and and the the complex complex QRST QRST can can be be easily easily The while the P wave isthe often confused with similarwith artefacts. identified, while while the wave is is often often confused confused with similar artefacts. artefacts. identified, PP wave similar

Sensors 2017, 17, 2474

9 of 10

4. Conclusions The motivations of this research activity can be summarized by the following needs: simple procedure, portability, remote monitoring, and wearable solutions. These objectives can be achieved providing a low number of electrodes, low-power, and size reduction. In the last eight years almost 20 papers have been published on this subject; 45% of them present devices based on capacitive coupling. One third of these studies are only simulated. In this paper the complete design and development process of a device for the contactless detection of ECG-like signals is presented, starting from the design simulation and ending with the fabrication of the device and the experimental tests. In this work, we presented a system for the contactless monitoring of human heart activity. This system has been designed to detect an ECG-like signal in non-contact mode, meaning not in direct contact with the skin, but at a certain distance, represented by the typical thickness of a basic cotton T-shirt. The electronic circuits composing the system were simulated and the prototype was tested in a real experiment on control subjects. The ECG signals registered account for an acceptable similarity with the ones obtained by the standard ECG monitoring instruments. The development of such a system opens the way for developing applications in different fields supported by distributed IoT (Internet of Things) data transfer systems such as space, industrial, and healthcare. For medical applications, the role of new wearable technologies will ensure a wider use of pre-screening, reducing the secondary effects of cardiac pathology and limiting medical care costs. On the other hand, for space applications it will allow the monitoring of the activities of astronauts without using specific and complex medical protocols. Further improvements will include the embedding of the device into some daily life households, such as in beds, chair and walls. Acknowledgments: This work was funded by the Italian Space Agency in the frame of the project: COR – Cardio Osservazione Remota. Thanks to Simone Grasso (University Campus Bio-Medico di Roma), Michele Tripoli and Ivano Musso (ALTEC S.p.A.). Author Contributions: F.R.P. fabricated the prototype; performed the experiment, wrote the paper; M.S. conceived and designed the system; performed the experiments; analyzed the data; A.Z. fabricated the prototype; performed the experiments; analyzed the data; M.B. conceived and designed the experiments; G.F. Desinged the electronic interface; wrote the paper; A.D. conceived and designed the system; analyzed the data; wrote the paper; G.P. conceived and designed the system; performed the experiment; wrote the paper. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2.

3. 4.

5. 6.

7. 8.

Da He, D.; Winokur, E.S.; Sodini, C.G. An ear-worn vital signs monitor. IEEE Trans. Biomed. Eng. 2015, 62, 2547–2552. Stornelli, V.; Ferri, G. A single current conveyor-based low voltage low power bootstrap circuit for ElectroCardioGraphy and ElectroEncephaloGraphy acquisition systems. Analog Integr. Circuit Signal Process. 2014, 79, 171–175. [CrossRef] Freeman, D.K.; Gatzke, R.D.; Mallas, G.; Chen, Y.; Brouse, C.J. Saturation of the Right-Leg Drive Amplifier in Low-Voltage ECG Monitors. IEEE Trans. Biomed. Eng. 2015, 62, 323–330. [CrossRef] [PubMed] Barone, U.; Merletti, R. Design of a portable, intrinsically safe multichannel acquisition system for high-resolution, real-time processing HD-Semg. IEEE Trans. Biomed. Eng. 2013, 60, 2242–2252. [CrossRef] [PubMed] Webster, J.G. Biopotential Amplifiers. In Medical Instrumentation Application and Design, 4th ed.; Wiley: Hoboken, NJ, USA, 2009; pp. 258–260. Wang, T.; Shen, H.P.; Lin, C.H.; Ou, Y.L. Detection and prediction of Sudden Cardiac Death (SCD) for Personal Healthcare. In Proceedings of the 2007 3rd International IEEE/EMBS Conference on Neural Engineering, Piscataway, NJ, USA, 2–5 May 2007; pp. 2575–2578. Neuman, M.R. Biopotential amplifiers. In Medical Instrumentation: Application and Design; Webster, J.G., Ed.; Wiley: New York, NY, USA, 1998; Chapter 6; pp. 256–258. Thakor, N.V. Biopotentials and Electro-physiology Measurement; Johns Hopkins School of Medicine: Baltimore, MD, USA, 1999.

Sensors 2017, 17, 2474

9.

10. 11. 12. 13. 14.

15. 16. 17. 18.

19. 20. 21. 22. 23.

24. 25. 26. 27. 28.

10 of 10

Prutchi, D.; Norris, M. Bandpass selection for biopotential amplifiers. In Design and Development of Medical Electronic Instrumentation: A Practical Perspective of the Design, Construction, and Test of Medical Devices, 1st ed.; Wiley: Hoboken, NJ, USA, 2005; pp. 2–6. Miller, H.A.; Harrison, D.C. Biomedical Electrode Technology; Academic: New York, NY, USA, 1974. Searle, A.; Kirkup, L. A direct comparison of wet, dry and insulating bioelectric recording electrodes. Physiol. Meas. 2000, 21, 271–283. [CrossRef] [PubMed] Burke, M.J.; Gleeson, D.T. A micropower dry-electrode ECG preamplifier. IEEE Trans. Biomed. Eng. 2000, 47, 155–162. [CrossRef] [PubMed] Lin, C.T.; Liao, L.D.; Liu, Y.H.; Wang, I.J.; Lin, B.S.; Chang, J.Y. Novel dry polymer foam electrodes for long-term EEG measurement. IEEE Trans. Biomed. Eng. 2011, 58, 1200–1207. [PubMed] Chiou, J.C.; Ko, L.W.; Lin, C.T.; Hong, C.T.; Jung, T.P.; Liang, S.F.; Jeng, J.L. Using novel MEMS EEG sensors in detecting drowsiness application. In Proceedings of the IEEE Biomedical Circuits and Systems Conference, San Diego, CA, USA, 13 November 2006; pp. 33–36. Chi, Y.M.; Jung, T.P.; Cauwenberghs, G. Dry-Contact and Noncontact Biopotential Electrodes: Methodological Review. IEEE Rev. Biomed. Eng. 2010, 3, 106–119. [CrossRef] [PubMed] Teichmann, D.; Foussier, J.; Jia, J.; Leonhardt, S.; Walter, M. Noncontact monitoring of cardiorespiratory activity by electromagnetic coupling. IEEE Trans. Biomed. Eng. 2013, 60, 2142–2152. [CrossRef] [PubMed] Sullivan, T.J.; Deiss, S.R.; Cauwenberghs, G. A low-noise, non-contact EEG/ECG sensor. In Proceedings of the Biomedical Circuits and Systems Conference, Montreal, QC, Canada, 27–30 November 2007; pp. 154–157. Ueno, A.; Akabane, Y.; Kato, T.; Hoshino, H.; Kataoka, S.; Ishiyama, Y. Capacitive sensing of electrocardiographic potential through cloth from the dorsal surface of the body in a supine position: A preliminary study. IEEE Trans. Biomed. Eng. 2007, 54, 759–766. [CrossRef] [PubMed] Lim, Y.G.; Kim, K.K.; Park, K.S. ECG recording on a bed during sleep without direct skin-contact. IEEE Trans. Biomed. Eng. 2007, 54, 718–725. [CrossRef] [PubMed] Wartzek, T.; Eilebrecht, B.; Lem, J.; Lindner, H.J.; Leonhardt, S.; Walter, M. ECG on the road: Robust and unobtrusive estimation of heart rate. IEEE Trans. Biomed. Eng. 2011, 58, 3112–3120. [CrossRef] [PubMed] Gao, S.; Huang, D.Y.; Fang, W.C. A non-contact ECG sensing system with motion artifact equalization. In Proceedings of the IEEE BioCAS, Shanghai, China, 17–19 October 2016; pp. 244–247. Kranjec, J.; Begus, S.; Drnovsek, J.; Gersak, G. Novel methods for noncontact heart rate measurement: A feasibility study. IEEE Trans. Instrum. Meas. 2014, 63, 838–847. [CrossRef] Santonico, M.; Zompanti, A.; Pennazza, G.; Ferri, G.; Parente, F.R.; Benassai, M. Contactless detection of ECG Signals: Sensor Architecture and Simulation. In Proceedings of the IEEE Computer Society Signature Conference on Computers, Software and Applications, Turin, Italy, 4–8 July 2017. Texas Instruments, Dallas, TX, USA. Datasheet, LF411, Low Offset, Low Drift JFET Input Operational Amplifier. Available online: http://www.ti.com/lit/ds/symlink/lf411-n.pdf (accessed on 4 July 2017). Baxter, L.K. Capacitive sensor basics. In Capacitive Sensors: Design and Applications; The Institute of Electrical and Electronics Engineers: New York, NY, USA, 1997; pp. 1–46. Ebrahim, N.; Deen, M.J.; Mondal, T. A wireless wearable ECG sensor for long-term applications. IEEE Commun. Mag. 2012, 50. [CrossRef] Sandra, K.R.; Anusha, A.S.; Mohan, N.M.; George, B. Simulation study of a contactless, capacitive ECG system. In Proceedings of the TENCON–2014 IEEE Region 10 Conference, Bangkok, Thailand, 22–25 October 2014. Texas Instruments, Dallas, TX, USA. Datasheet, INA128, Low Power Instrumentation Amplifier. Available online: http://www.ti.com/lit/ds/sbos051c/sbos051c.pdf (accessed on 4 July 2017). © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).