Ultra-Wide Patch Antenna Array Design at 60 GHz Band for Remote Vital Sign Monitoring with Doppler Radar Principle

Transcription

1 J Infrred Milli Terhz Wves (217) 38:48 66 DOI 1.17/s z Ultr-Wide Ptch Antenn Arry Design t 6 GHz Bnd for Remote Vitl Sign Monitoring with Doppler Rdr Principle Muhmmd Sqi Rni 1 & Hooshng Ghfouri-Shirz 1 Received: 11 July 216 / Accepted: 18 Novemer 216 / Pulished online: 2 Decemer 216 # The Author(s) 216. This rticle is pulished with open ccess t Springerlink.com Astrct In this pper, ultr-wide ptch ntenn rrys hve een presented t 6 GHz nd ( GHz) with improved gin nd em-width cpilities for remote detection of respirtion nd hert et rte of person with Doppler rdr principle. The ntenns mesured nd simultion results showed close greement. The rething rte (BR) nd hert rte (HR) of 31-yer-old mn hve een ccurtely detected from vrious distnces rnging from to 2 cm with oth single-ntenn nd dul-ntenn opertions. In the cse of single-ntenn opertion, the signl is trnsmitted nd received with the sme ntenn, wheres in dulntenn opertion, two identicl ntenns re employed, one for signl trnsmission nd the other for reception. It hs een found tht in cse of the single-ntenn opertion, the ccurcy of the remote vitl sign monitoring (RVSM) is good for short distnce; however, in the cse of the dul-ntenn opertions, the RVSM cn e ccurtely crried out t reltively much longer distnce. On the other hnd, it hs lso een seen tht the visul results re more ovious with higher gin ntenns when the rdr em is confined just on the suject s ody re. Keywords Microstrip ntenn. Ultr-wideptch rrys. Remote vitl sign monitoring. Doppler rdr. Non-contct vitl signs monitoring. 6 GHz ntenn rrys 1 Introduction Non-contct detection of respirtion nd hert et rte with Doppler rdr is more convenient wy to check the vitlity signs of person s compred to the conventionl * Hooshng Ghfouri-Shirz ghfourh@hm.c.uk Muhmmd Sqi Rni sqirni@hotmil.com 1 School of Electronic Electricl nd System Engineering, University of Birminghm, Edgston B1 2TT, UK

2 J Infrred Milli Terhz Wves (217) 38: vitl sign monitoring devices, such s electrocrdiogrm (ECG), pulse oximetry nd cpnogrphy, ecuse the conventionl devices need direct sensors plnttion on the suject ody s well s they consume comprtively much longer implementtion time to strt the mesurements [1 8]. However, the results ccurcy nd reliility of RVSM system need to e delt ppropritely for its rel-life pplictions [2]. RVSM finds vst numer of pplictions in regulr nd specil helth cre, emergency services, security nd defence sectors [1 12]. In the recent couple of decdes, significnt mount of reserch hs een conducted to improve the ccurcy of the RVSM y employing vrious techniques including distnce nd frequency optimiztion [13 18], roust signl processing methods [19 21], reth holding for HR detection nd so on [22 28]. However, most of the reported prcticl work is sed on Doppler rdrs working t lower microwve (MW) frequency nds round 2.4,.8 nd 1 GHz where the wvelength resolution of the electromgnetic wve is low nd therefore the ccurcy of the vitl signs detection, especilly the HR, remined chllenging [23]. Secondly, the rdr ntenn systems t the mentioned low MW frequencies re quite ulky which my stin to e integrted with modern rel-life compct devices like smrt phones nd tlets [23]. K nd frequencies hve lso een investigted for RVSM to enhnce the RVSM sensors sensitivity due to comprtively shorter wvelength t K frequencies [24 26]. In the recent yers, millimetre-wve (MMW) frequencies (3 3 GHz) hve drwn remrkle reserch interest for RVSM implictions [23, 27 3]. The min motives of RVSM t MMW frequencies include (i) improvement in the detection ccurcy y employing shorter wvelength of the signl, (ii) smller form fctor for the device compctness nd (iii) possiility of more suject focused signl trnsmission nd reception to void the interference from the unwnted side reflections [22, 33 3]. From MMW nd, 6 GHz nd (7 66 GHz) hs drwn more ttrction ecuse this nd is free of licence nd is quite mture due to its extensive use for severl other wireless services [27]. Antenn designing for Doppler rdr plys crucil role in the precision of RVSM. The ntenn should e designed in such wy tht it only focuses the rdition em on the suject [22] nd should hve n dequte gin to mintin required signl to noise rtio [31]. Microstrip ptch ntenns re thought to etter option for compct MMW sensors due to their emedment with on-chip devices, low profile, low cost nd ility to mke rry to ttin high ntenn gin [23, 3 32]. In [3], conventionlly designed microstrip ptch ntenns re integrted with on-chip micro-rdr system for RVSM. However, the nrrow microstrip trnsmission line (TL) designed for impednce mtching nd feeding purpose is fricted with flip-chip method. In [31], two circulrly polrised ptch ntenn rrys of elements hve een presented for vitl sign detection t GHz. The ntenn rrys nd the ssocited nrrow feed lines were fricted y using specil technique sed on multilyer low-temperture co-fire cermic (LTCC) sustrte. The ntenns gin remined s low s 4.86 dbi in cse of two-element nd 9.7 dbi in four-element rrys. Furthermore, for RVSM with these rrys, the suject hs to hold rething for while for ccurte detection of HR. Therefore, regrding RVSM t 6 GHz, there re still some gps in the literture, i.e. cost-effective microstrip ntenn rrys designed for ccurte nd simultneous detection of BR nd HR nd the study of RVSM with single ntenn for very short distnce nd with doule ntenns for long distnces.

3 J Infrred Milli Terhz Wves (217) 38:48 66 In the present pper, three microstrip ntenn rrys of ultr-wide elements hve een designed for ccurte detection of humn BR nd HR t 6 GHz nd frequencies. The use of rry ntenn is mde to minimise the interferences of the reflected signls from the side Fig. 1 Structures of the designed ntenn rrys with () 2 1,() 3 2nd(c) 6 2 ptch elements

4 J Infrred Milli Terhz Wves (217) 38: Fig. 2 Fricted ntenn rrys with ()-() 2 1,(c)-(d) 3 2 nd(e)-(f) 6 2 ptch elements ojects ecuse the rry ntenns hve nrrower rdition emwidths thn the single ptch ntenn [36]. Suitility of 6 GHz nd for RVSM hs een studied under the Doppler rdr principle. Both BR nd HR hve een simultneously mesured of the suject sitting in front of the ntenns nd hving norml rething. The detection process hs een studied in two wys: (i) with single ntenn used for oth trnsmission nd reception nd (ii) with dul ntenns, one for trnsmission nd the other for reception. The ultr-wide ptch elements provided good trde-off etween the rrys gin nd size. The low side-lo levels, high gins, nd nrrow em widths of the rrys confined the EM wve on the suject chest nd ultimtely improved the ccurcy of RVSM. Moreover, ll of the rry dimensions were wide enough for conventionl cost-effective friction with the ordinry PCB etching technology. 2 Antenn Design Three microstrip ptch ntenn rrys hve een designed t 6 GHz nd on low loss Duroid sustrte with thickness h =24μm nd dielectric constnt ε r = 2.2. The structure digrms of the rrys re presented in Fig. 1 where the rrys in Fig. 1 c re comprised of 2 1, 3 2 nd 6 2 ptch elements, respectively. The ptch width (W) nd length (L) re clculted s [37]: W ¼ λ ð2m þ 1Þ rffiffiffiffiffiffiffiffiffiffiffiffiffi ε r þ ð1þ L ¼ λ gð2n þ 1Þ p 2 ffiffiffiffiffiffiffiffi 2ΔL ð2þ ε reff Where M nd N re non-negtive integers (in the current cse M =1ndN =)ndλ nd λ g re free spce nd guided wvelengths, respectively. ε r nd ε reff re reltive nd effective Tle 1 Dimensions (mm) of the fricted rrys L W L T1 L T2 W T1 W T2 S Ant. 2() Ant. 2(c) Ant. 2(e)

5 2 J Infrred Milli Terhz Wves (217) 38:48 66 dielectric constnts, respectively. ΔL is the ptch length extension due to the fringing field effect [38]. The specified ptch width extension (i.e. M = 1in Eq. (1)) is dopted to improve the rrys gin y round 3 db s compred to the gin of n rry with conventionl ptch width (i.e. M =)[37]. Secondly, y employing the ultr-wide ptches, the ntenn s input impednce is Fig. 3 Mesured nd simultion S 11 of ntenn rrys shown () in Fig. 2()-(), () in Fig. 2(c)-(d) nd (c) in Fig. 2(e)-(f) -1 Experiment Results S11 (db) -2-3 Mesured -() Mesured -() Simultion Frequency (GHz) Experiment Results -1 S11 (db) -2-3 Mesured -(c) Mesured -(d) Simultion Frequency (GHz) c Experiment Results S11 (db) Mesured -(e) Mesured -(f) Simultion Frequency (GHz)

6 J Infrred Milli Terhz Wves (217) 38: Fig. 4 Simultion 3-D FFPs of ntenn rrys shown () in Fig. 2()-(), () in Fig. 2(c)-(d) nd (c) in Fig. 2(e)-(f)

7 4 J Infrred Milli Terhz Wves (217) 38:48 66 reduced which ultimtely helps to widen the feeding TL nd hence to improve the friction tolernce should the conventionl low cost PCB etching method e employed [37]. The TL length L T1 is deployed s impednce trnsformer which is given s L T1 =(2P + 1) λ/4, where P is non-negtive integer (in the current cse P = 1). The inter-seprtions etween the series nd prllel ptch elements re set to e L T2 ¼ λ g =2 þ 2ΔL nd S = λ g /2, respectively (see Fig. 1), to mtch the current phse on the ptch elements. The TL width (W T1 ) is computed sed on the totl input impednce (Z )oftherrywithn numer of symmetricl ptch elements [37]. A generl expression for Z is Z ¼ 9:81λ nw Z is mtched with the stndrd Ω impednce through TL L T1 with the chrcteristic p impednce Z 1 ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Z. The TL width (W T2 ) is computed sed on the totl input impednce (Z 2 )oftheseries ptch elements in line on one side of the rry. The chrcteristic impednce of the line width W T2 is given s Z 2 =9.81λ /(mw), where m is numer of ptch elements in series. The following is the generl expression used to clculte the TL width W T for the chrcteristic impednce Z [39]: ð3þ W T ¼ 7:47 h e x 1:2t ð4þ Fig. Mesured nd simultion FFP of ntenn rrys shown in Fig. 2() nd in Fig. 2() Mgnitude(dB) Fr-Field Pttern -2-3 Mesured-H-plne Simultion-H-plne -3 Mesured-E-plne Simultion-E-plne Angle(deg.) Mgnitude(dB) Fr-Field Pttern Mesured-H-plne -3 Simultion-H-plne -3 Mesured-E-plne Simultion-E-plne Angle(deg.)

8 J Infrred Milli Terhz Wves (217) 38:48 66 p where x ¼ Z ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ε r þ 1:41=87 nd t is copper cldding (see Fig. 1).The finl dimensions of the rrys re optimised with CST microwve studio for the est ntenn performnce. 3 Antenn Arrys Results Figure 2 demonstrtes the fricted ntenn rrys where ()-(), (c)-(d) nd (e)-(f) re two copies of the rry with 2 1, 3 2 nd 6 2 numer of ptch elements, respectively. Tle 1 shows the dimensions of the fricted rrys where it cn e seen tht ll of the rrys dimensions re well ove the PCB etching limit of 12-μm line width/gp. Figure 3 c illustrte the simultion nd mesured return loss responses of the rrys shown in Fig. 2()-(), (c)-(d) nd (e)-(f). As it cn e seen from Fig. 3, the resonnces of ntenns 2()-() re centred t round 63.4 GHz nd tht of ntenns 2(c)-(f) re centred in the rnge of 64. to 6 GHz ut re etter thn 1 db t 64.8 GHz. Therefore, 63.4 nd 64.8 GHz frequency tones hve een respectively selected for the RVSM when ntenns 2()-() nd 2(c)-(f) re employed. Figure 4 c present the simulted fr-field ptterns (FFP) in 3-D of the ntenns shown in Fig. 2()-(), (c)-(d) nd(e)-(f), respectively, wheres Figs., 6 nd 7 respectively show their simultion nd mesured fr-field ptterns (FFP) in H nd E plnes t their selected frequencies s mentioned in the lst prgrph. All of the ntenns mesured S 11 nd FFP results re within the PCB friction limit of % [31]. Some distortions in the ntenns FFP in Figs. 6 nd 7 cn e oserved which re due to the reflections from the surroundings nd test fixtures [4]. It cn Fig. 6 Mesured nd simultion FFP of ntenn rrys shown in Fig. 2(c) nd in Fig. 2(d) Mgnitude(dB) Fr-Field Pttern -3 Mesured-H-plne Simultion-H-plne -3 Mesured-E-plne Simultion-E-plne Angle(deg.) Mgnitude(dB) Fr-Field Pttern -2 Mesured-H-plne -3 Simultion-H-plne -3 Mesured-E-plne Simultion-E-plne Angle(deg.)

9 6 J Infrred Milli Terhz Wves (217) 38:48 66 Fig. 7 Mesured nd simultion FFP of ntenn rrys shown in Fig. 2(e) nd in Fig. 2(f) Mgnitude(dB) Fr-Field Pttern Mesured-H-plne -3 Simultion-H-plne -4 Mesured-E-plne Simultion-E-plne Angle(deg.) Mgnitude(dB) Fr-Field Pttern -3 Mesured-H-plne -3 Simultion-H-plne -4 Mesured-E-plne Simultion-E-plne Angle(deg.) e noticed from Figs. 4,, 6 nd 7 tht the ntenns min los in the rdition ptterns re directed lmost towrds in oth E nd H plnes for ll the cses. The ntenns mesured 1 db return loss ndwidths (BW) nd gins re summrised in Tle 2 where it is ovious tht gin of the rrys hs significntly improved when more ptch elements hve used ut the BW hs reduced. However, the BW is not crucil for our current ppliction of RVSM ecuse only single tone frequency will e used for the detection process. 4 Applicility of 6 GHz Doppler Rdr for RVSM Figure 8 shows the lock digrm of vrious stges of RVSM process when Doppler rdr principle is used for the vitl signs detection. A single frequency tone of continuous electromgnetic wve (CW) is trnsmitted through trnsmitter (T x ) ntenn. The wve is reflected ck from the suject chest locted t certin distnce d nd is received y receiver (R x ) ntenn. The qusi-periodic virtion of chest due to respirtion nd het et is phse Tle 2 Performnce of ntenn rrys shown in Fig. 2()-(f) Ant. () () (c) (d) (e) (f) BW (GHz) Gin (dbi)

10 J Infrred Milli Terhz Wves (217) 38: Fig. 8 Block digrm of RVSM process modulted on the received signl. This phse-modulted signl t the R x is correlted with trnsmitted signl nd the outcome dt is recorder for certin period of time. Susequently, the recorded rw dt which is in time domin is processed through vrious signl processing techniques, i.e. digitl filtering nd Fourier Trnsformtion, to extrct the respirtion nd hert et rte of the suject. According to the Doppler rdr theory, for trnsmitted signl S(t)= cos(2πft + φ(t)), where f nd φ(t) re the frequency nd phse noise of the trnsmitted wve, respectively, the received se nd signl R(t) my e pproximted s [41]: Rt ðþ¼cos θðþþ t 4πx ðþ t þ 4πx hðþ t ðþ λ λ where θ(t) is the totl phse shift due to the signl pth (d), reflections from the suject nd surroundings nd residul phse noise. λ, x (t) ndx h (t) re the operting wvelength, chest virtion displcement due to respirtion nd hertet, respectively. Due to the periodic nture of the x (t) ndx h (t), they my e pproximted s x (t)=m sin(2πf t)ndx h (t)=m h sin(2πf h t), where m nd m h re the displcement mplitudes of the chest motion due to respirtion nd hertet, respectively. f nd f h re the frequencies of BR nd HR, respectively. This wy, the expnsion of Eq. (4) in Fourier series leds to [24]: Rt ðþ¼ X X J j i¼ j¼ 4πm λ 4πm h cosðj 2πf λ t þ i 2πf h t þ θþ ð6þ J i Where J n (X) is Bessel function of first kind with rgument X. Tking the first positive hrmonics of oth f nd f h into ccount, the ove eqution cn e written s 4πm 4πm h Rt ðþ¼j 1 J cosð2πf λ λ t þ θþ ð7þ 4πm 4πm h þ J J 1 cosð2πf λ λ h t þ θþ 4πm where J 1 λ J 4πm h λ nd J 4πm λ J 4πm h 1 λ re the mplitudes of the phse vritions in Fig. 9 Plots of Bessel functions J (X), J 1 (X) nd their product J (X) J 1 (X) for rgument X. Mgnitude of the phse mplitude of R(t)inEq.(6) due to oth BR nd HR in comined, c just BR signl nd d just HR signl, for vrious comintions of m nd m h

11 8 J Infrred Milli Terhz Wves (217) 38:48 66 Bessel functions of first kind for n =,1 1 J J n (X). J 1 J *J X Phse vrition (deg.) due to BR & HR Chest displcement (mm) with HR Chest displcement (mm) with BR 12 c Phse vrition (deg.) due to BR Chest displcement (mm) with HR Chest displcement (mm) with BR d Phse vrition (deg.) due to HR Chest displcement (mm) with HR Chest displcement (mm) with BR

12 J Infrred Milli Terhz Wves (217) 38: Fig. 1 Experiment setup to mesure the RVSM t 6 GHz nd R(t) due to respirtion nd hertet, respectively. Eqution (6) contins the essentil informtion relted to the pplicility of 6 GHz nd frequencies for RVSM. Bsed on Eq. (6), we re going to nlyse nd visulise the key fctors involved in RVSM t 6 GHz nd. Figure 9 shows the plots of J (X), J 1 (X) nd their product for the rgument X up to 3 where it cn e noticed tht J (X) ndj 1 (X) re the periodic functions with some phse shift nd the mximum mplitude of their product is much smller thn their individul mximum mplitudes nd it psses through null vlues twice s compred to individul J (X) ndj 1 (X) curves. Furthermore, the mplitude of J (X) J 1 (X) diminishes for higher vlues of rgument X.NoweringEq.(6) in mind, the mplitude of the received BR nd HR signl is comprised of the product of J (X) ndj 1 (X) ndthergumentx is controlled y m, m h nd λ. For our current cse of VSM t 6 GHz nd, λ is round 4.6 mm, m =(8 12)mm nd m r =(.2.)mm for person t rest with norml rething [42]. Figure 9 d shows the phse mplitude vritions of R(t) signl for vrious comintion of m nd m r with BR nd HR in comined, only BR nd HR, respectively. From Fig. 9, it cn e seen tht the mximum R(t) mplitude goes to round 4 which is high enough for RVSM detection; however, there is null detection line t chest mplitude of round 9.9 mm. Figure 9c indictes similr null Fig. 11 RVSM mesurement with ntenn shown in Fig. 2(). Recorded rw dt in time domin. Detected BR nd HR peks S11 phse (deg.) S11 phse (deg.)

13 6 J Infrred Milli Terhz Wves (217) 38:48 66 Fig. 12 RVSM mesurement with ntenn shown in Fig. 2(c). Recorded rw dt in time domin. Detected BR nd HR peks S11 phse (deg.) S11 phse (deg.) line position for just BR signl. On the other hnd, Fig. 9d shows tht in cse of the received HR signl, there re two null detection lines t chest displcements of 8.8 nd 11 mm nd the mximum phse mplitude of hertet is out one third of tht of respirtion. However, the proility of these undesired m nd m h comintions nd hence the null detection points is very low nd smll vrition in λ cn e mde to void these points (see Eq. (6)) [24]. As conclusion so fr, oth BR nd HR signls hve high proility of detection with RVSM t 6 GHz nd ut the received BR signl my e clerer thn HR signl. RVSM Results nd Discussions Figure 1 shows the experiment setup to mesure the RVSM t 6 GHz nd. The 67 GHz Rohde nd Schwrz VNA is used s trnsmitter nd receiver. The three ntenn designs s presented in section III (see Fig. 2) hve een employed one y one for signl trnsmission nd reception. As mentioned in section I, for ech ntenn design, the RVSM dt hs een cquired with two wys; (i) with single ntenn employed for oth trnsmission nd reception nd (ii) with doule ntenns (two similr ntenns), one for trnsmission nd the other for reception. A normlly rething person sits in front of the ntenns itertively t different distnces (d) nd the RVSM dt is recorded for 6 s for ech itertion. The recorded dt is then processed in Mtl progrmme through vrious digitl signl processing techniques nd the trgeted BR nd HR hve een extrcted. The mjor prts of digitl signl Fig. 13 RVSM mesurement with ntenn shown in Fig. 2(e). Recorded rw dt in time domin. Detected BR nd HR peks S11 phse (deg.) S11 phse (deg.)

14 J Infrred Milli Terhz Wves (217) 38: Tle 3 Mesured BR nd HR with single ntenn in Fig. 2 Antenn 2() 2(c) 2(e) BR HR processing include digitl filtering of the recorded rw dt in time domin nd Discrete Fst Fourier Trnsformtion (DFFT) of the filtered dt [43 46]..1 RVSM with Single Antenn As mentioned erlier, in the single-ntenn opertion, only one of ech designed ntenn rrys is employed for oth signl trnsmission nd reception nd the vitl signs cn only e ccurtely detected for very short distnce [6]. The ntenn is plced t out cm wy from the suject chest nd the phse of the reflection coefficients (S 11 ), which contin the HR nd BR informtion, hs een recorded for 6 s. Figs. 11, 12 nd 13 illustrte the mesured RVSM dt with the individul-ntenn rrys shown in 2(), (c) nd (e), respectively. Figs. 11, 12 nd 13 represent the recorded rw dt of S11 phse nd Figs. 11, 12 nd 13 depict the DFFT of the processed dt in frequency (1/min) domin. As seen from Figs. 11, 12 nd 13, in ech cse, the first nd the second highest peks represent the detected BR nd HR, respectively. All of the detected BR nd HR re listed in Tle 3 where it is ovious tht oth BR nd HR re consistent within their respective rnges for the ordinry suject (31-yer-old mn) [47]. The mesured BR nd HR results re lso mtched with the results otined with mnul counting y using stopwtch..2 RVSM with Doule Antenns In RVSM with doule ntenn opertion, two similr ntenns re deployed for vitl sign detection, one for trnsmitting the EM wve nd the other for receiving the reflected wve. Both ntenns re fixed t n equl distnce from the suject chest with their min ems directed to the chest. The inter-seprtion etween the ntenns is set to e out 2 cm. The phse of S21 is recorded for 6 s for RVSM with ech ntenn design shown in Fig. 2()-(f). The ccurcy of the RVSM hs een studied for vrious distnces (d) etween the suject nd ntenns rnging from 2 to 2 cm. Figures 14, 1 nd 16 demonstrte the RVSM dt with ntenn design 2()-() from distnce.2, 1 nd 2 m, respectively. Similrly, Figs. 17 nd 18 Fig. 14 RVSM mesurement with ntenn shown in Fig. 2()-() from.2 m. Recorded rw dt in time domin. Detected BR nd HR peks phse (deg.) phse (deg.) S

15 62 J Infrred Milli Terhz Wves (217) 38:48 66 Fig. 1 RVSM mesurement with ntenn shown in Fig. 2()-() from 1 m. Recorded rw dt in time domin. Detected BR nd HR peks S21 phse (deg.) S21 phse (deg.) Fig. 16 RVSM mesurement with ntenn shown in Fig. 2()-() from 2 m. Recorded rw dt in time domin. Detected BR nd HR peks S21 phse (deg.) S21 phse (deg.) nd Figs. 19 nd 2 demonstrte the RVSM dt with ech ntenn design 2(c)-(d) nd 2(e)-(f) from distnce.2 nd 1 m, respectively. The su-figs. () nd () in Figs. 14, 1, 16, 17, 18, 19 nd 2 represent the recorded rw dt in time domin nd the processed dt in frequency (1/min) domin, respectively. The first nd the second pek in the processed dt in Figs. 14, 1, 16, 17, 18, 19 nd 2 represent the mesured BR nd HR, respectively, where it cn e seen tht oth BR nd HR peks re very ovious cross the frequency spectrum. The mesured BR nd HR results of RVSM with dul ntenns up to 1 m distnces re shown Fig. 17 RVSM mesurement with ntenn shown in Fig. 2(c)-(d) from.2 m. Recorded rw dt in time domin. Detected BR nd HR peks S21 phse (deg.) S21 phse (deg.)

16 J Infrred Milli Terhz Wves (217) 38: Fig. 18 RVSM mesurement with ntenn shown in Fig. 2(c)-(d) from 1 m. Recorded rw dt in time domin. Detected BR nd HR peks S21 phse (deg.) S21 phse (deg.) Fig. 19 RVSM mesurement with ntenn shown in Figs. 2(e)-(f) from.2 m. Recorded rw dt in time domin. Detected BR nd HR peks S21 phse (deg.) S21 phse (deg.) in Tle 4. The detected BR nd HR with ntenn design 2()-() in cse of 2 m distnce re 19/min nd 79/min, respectively. In ll the cses, oth BR nd HR results re consistent within their expected rnges nd re mtched with the results otined y mnul counting s mentioned erlier. Fig. 2 RVSM mesurement with ntenn shown in Fig. 2(e)-(f) from 1m. Recorded rw dt in time domin. Detected BR nd HR peks S21 phse (deg.) S21 phse (deg.)

17 64 J Infrred Milli Terhz Wves (217) 38:48 66 Tle 4 Mesured BR nd HR with dul ntenns Antenn 2()-() Antenn 2(c)-(d) Antenn 2(e)-(f) d BR-HR BR-HR BR-HR.2 m m Conclusion Three microstrip ntenn rrys hve een designed sed on ultr-wide ptch elements with improved gin nd rdition chrcteristics for ccurte RVSM t 6 GHz nd frequencies. The ntenn simultion nd mesured results greed well in ll cses. The fesiility of 6 GHz chnnel hs een studied for RVSM ppliction under the Doppler rdr principle. The RVSM mesurements hve een conducted with ech designed rry in (i) single-ntenn opertion, where the sme ntenn working s T x nd R x is plced t cm from the suject chest, s well s (ii) in doule ntenns opertion where two identicl ntenns, one for T x nd the other for R x, hve een deployed t vrious distnces from the suject to see the consistencyintheresultsccurcy.bothbrnd HR hve een detected for the suject with norml rething throughout the dt recording time nd no reth holding ws required for ccurte HR detection. In ll of the cses, the mesured results showed very ccurte, cler nd roust detection of RVSM with simple digitl signl processing techniques. For future work, the proposed ntenn rrys could e used for outdoor RVSM detection from even longer distnce. Furthermore, the potentil error sources like the signl reflections from the side ojects through the ntenns side loes, ccurcy of the vitl signs of the diversity of ptient types cn e investigted in detils. Open Access This rticle is distriuted under the terms of the Cretive Commons Attriution 4. Interntionl License ( which permits unrestricted use, distriution, nd reproduction in ny medium, provided you give pproprite credit to the originl uthor(s) nd the source, provide link to the Cretive Commons license, nd indicte if chnges were mde. References 1. L. Sun, Y. Li, H. Hong, F. Xi, W. Ci, X. Zhu, Super-resolution spectrl estimtion in short-time non-contct vitl sign mesurement, Review of Scientific Instruments, 21 Apr 1;86(4): A. D. Droitcour, T. B. Seto, B-K. Prk, S. Ymd, A. Vergr, C. E. Hourni, T. Shing, A.Yuen, V. M. Luecke, nd O. Boric-Luecke, Non-contct respirtory rte mesurement vlidtion for hospitlized ptients, In 29 Annul Interntionl Conference of the IEEE Engineering in Medicine nd Biology Society, pp IEEE, (29) 3. M. Villrroel, A. Guzzi, J. Jorge, S. Dvis, P. Wtkinson, G. Green, A. Shenvi, K. McCormick, nd L. Trssenko, Continuous non-contct vitl sign monitoring in neontl intensive cre unit, Helthcre technology letters 1, no. 3 (214): S. Suzuki, T. Mtsui, M. Kgw, T. Aso, nd K. Kotni, An pproch to non-contct vitl sign monitoring using dul-frequency microwve rdrs for elderly cre, Journl of Biomedicl Science nd Engineering 6, no. 7 (213): 74. M. Uenoym, T. Mtsui, K. Ymd, S. Suzuki, B. Tkse, S. Suzuki, M. Ishihr, nd M. Kwkmi, Non-contct respirtory monitoring system using ceiling-ttched microwve ntenn, Medicl nd Biologicl Engineering nd Computing 44, no. 9 (26): 83-84

18 J Infrred Milli Terhz Wves (217) 38: D. Oeid, G. Zhri, S. Sdek nd G. El Zein, 212. Microwve doppler rdr for hertet detection vs electrocrdiogrm. Microwve nd Opticl Technology Letters, 4(11), pp Ko, T.Y.J. nd Lin, J., 213, April. Vitl sign detection using 6-GHz Doppler rdr system. In Wireless Symposium (IWS), 213 I.E. Interntionl (pp. 1-4). IEEE. 8. C. Gu, Short-Rnge Noncontct Sensors for Helthcre nd Other Emerging Applictions: A Review. Sensors, 16(8), 216, p L. Ren, Y. S. Koo, H. Wng, Y. Wng, Q. Liu, A. E. Fthy, Noncontct Multiple Hertets Detection nd Suject Locliztion Using UWB Impulse Doppler Rdr, IEEE Microwve nd Wireless Components Letters 2, no. 1 (21): F. Adi, H. Mo, Z. Kelc, D. Kti nd R. C. Miller, Smrt homes tht monitor rething nd hert rte, In Proceedings of the 33rd Annul ACM Conference on Humn Fctors in Computing Systems, pp ACM, (21) 11. S. Suzuki, T. Mtsui, H. Kwhr, H. Ichiki, J. Shimizu, Y. Kondo, S. Gotoh, H. Yur, B. Tkse nd M. Ishihr, A non-contct vitl sign monitoring system for mulnces using dul-frequency microwve rdrs, Medicl & iologicl engineering & computing 47, no. 1 (29): C. Brüser, C. H. Antink, T. Wrtzek, M. Wlter, nd S. Leonhrdt, Amient nd Unotrusive Crdiorespirtory Monitoring Techniques, IEEE reviews in iomedicl engineering 8 (21): C. Li, Y. Xio nd J. Lin, A GHz Doule-Sidend Rdr Sensor Chip in.18 m CMOS for Non-Contct Vitl Sign Detection, IEEE Microwve nd Wireless Components Letters 18, no. 7 (28): S. Kzemi, A. Ghorni, H. Amindvr nd D. R. Morgn, Vitl-Sign Extrction Using Bootstrp-Bsed Generlized Wrlet Trnsform in Hert nd Respirtion Monitoring Rdr System, IEEE Trnsctions on Instrumenttion nd Mesurement 6, no. 2 (216): A.D.Droitcour,O.Boric-Luecke,V.M.Luecke,J.E.N.S.H.A.N.LinndG.T.A.Kovcs,Rnge correltion effect on ISM nd I/Q CMOS rdr for non-contct vitl signs sensing, In Microwve Symposium Digest, 23 I.E. MTT-S Interntionl, vol. 3, pp IEEE, (23) 16. M-C. Hung, J. J. Liu, W. Xu, C. Gu, C. Li, nd M. Srrfzdeh, A self-clirting rdr sensor system for mesuring vitl signs, IEEE trnsctions on iomedicl circuits nd systems 1, no. 2 (216): μm CMOS for non-contct vitl sign detection, In 29 I.E. Rdio Frequency Integrted Circuits Symposium, pp IEEE, (29) 18. C. Li, Y. Xio nd J. Lin, Design guidelines for rdio frequency non-contct vitl sign detection, In 27 29th Annul Interntionl Conference of the IEEE Engineering in Medicine nd Biology Society, pp IEEE, (27) 19. C. Li nd J. Lin, Complex signl demodultion nd rndom ody movement cncelltion techniques for non-contct vitl sign detection, In Microwve Symposium Digest, 28 I.E. MTT-S Interntionl, pp IEEE, (28) 2. C. Li, J. Ling, J. Li nd J. Lin, Accurte Doppler rdr noncontct vitl sign detection using the RELAX lgorithm, IEEE Trnsctions on Instrumenttion nd Mesurement 9, no. 3 (21): Rhmn, E. Yvri, X. Go, V. Luecke nd O. Boric-Luecke, Signl processing techniques for vitl sign monitoring using moile short rnge Doppler rdr, In Biomedicl Wireless Technologies, Networks, nd Sensing Systems (BioWireleSS), 21 I.E. Topicl Conference on, pp IEEE, (21) 22. C. Li, V. M. Luecke, O. Boric-Luecke nd J. Lin, A review on recent dvnces in Doppler rdr sensors for noncontct helthcre monitoring, IEEE Trnsctions on microwve theory nd techniques 61, no. (213): T-Y. J. Ko, A. Y-K. Chen, Y. Yn, T-M. Shen nd J. Lin, A flip-chip-pckged nd fully integrted 6 GHz CMOS micro-rdr sensor for hertet nd mechnicl virtion detections, In 212 I.E. Rdio Frequency Integrted Circuits Symposium, pp IEEE, (212) 24. C. Li nd J. Lin, Optiml crrier frequency of non-contct vitl sign detectors, In 27 I.E. Rdio nd Wireless Symposium, pp IEEE, (27) 2. C. Li nd J. Lin, Non-contct mesurement of periodic movements y 22-4GHz rdr sensor using nonliner phse modultion, In 27 IEEE/MTT-S Interntionl Microwve Symposium, pp IEEE, (27) 26. N. Birsn nd D-P. Muntenu, Non-contct crdiopulmonry monitoring lgorithm for 24 GHz Doppler rdr, In 212 Annul Interntionl Conference of the IEEE Engineering in Medicine nd Biology Society, pp IEEE, H-R. Chung, H-C. Kuo, F-L. Lin, T-H. Hung, C-S. Kuo nd Y-W. Ou, 6-GHz millimeter-wve life detection system (MLDS) for noncontct humn vitl-signl monitoring, IEEE Sensors Journl 12, no. 3 (212): H-C. Kuo, H-H. Wng, P-C. Wng, H-R. Chung, nd F-L. Lin, 6-GHz millimeter-wve life detection system with clutter cnceller for remote humn vitl-signl sensing, In Microwve Workshop Series on Millimeter Wve Integrtion Technologies (IMWS), 211 I.E. MTT-S Interntionl, pp IEEE, (211)

19 66 J Infrred Milli Terhz Wves (217) 38: H-C. Kuo nd H-R. Chung, Investigtion of crrier frequency effect on detection performnce of Doppler sensor systems for noncontct humn vitl-signs sensing, In 214 8th Interntionl Symposium on Medicl Informtion nd Communiction Technology (ISMICT), pp IEEE, (214) 3. T-Y. J. Ko, Y. Yn, T-M. Shen, A. Y-K. Chen nd J. Lin, Design nd nlysis of 6-GHz CMOS Doppler micro-rdr system-in-pckge for vitl-sign nd virtion detection, IEEE Trnsctions on Microwve Theory nd Techniques 61, no. 4 (213): T-M. Shen, T-Y. J. Ko, T-Y. Hung, J. Tu, J. Lin nd R-B. Wu, Antenn design of 6-GHz micro-rdr system-in-pckge for noncontct vitl sign detection, IEEE Antenns nd Wireless Propgtion Letters 11 (212): J. Go, K. Li, T. Sto, J. Wng, H. Hrd S. nd Kto, Jnury. Implementtion considertions of ptch ntenn rry for 6GHz em steering system pplictions. In 29 I.E. Rdio nd Wireless Symposium(pp. 3-38). IEEE, Bkhtiri, Ssn, Thoms W. Elmer, Nichols M. Cox, Nchpp Goplsmi, Appostolos C. Rptis, Sholin Lio, Ily Mikhelson, nd Aln V. Shkin. BCompct millimeter-wve sensor for remote monitoring of vitl signs.^ IEEE Trnsctions on Instrumenttion nd Mesurement 61, no. 3 (212): S. Bkhtiri, S. Lio, T. Elmer nd A. C. Rptis, A rel-time hert rte nlysis for remote millimeter wve IQ sensor, IEEE Trnsctions on Biomedicl Engineering 8, no. 6 (211): D. T. Petkie, C. Benton nd E. Bryn, Millimeter wve rdr for remote mesurement of vitl signs, (29): Yng, Z., Pthk, P.H., Zeng, Y., Lirn, X. nd Mohptr, P., Monitoring Vitl Signs Using Millimeter Wve. DOI: [Online] Aville t: [ com/files/mmvitl-moihoc.pdf], ccessed on Octoer the 1th, M. S. Rni nd H. Ghfouri-Shirz, Size improvement of rectngulr microstrip ptch ntenn t MMwve nd terhertz frequencies, Microwve nd Opticl Technology Letters 7, no. 11 (21): R. Bncroft, Microstrip nd printed ntenn design [electronic resource]. Rleigh, NC, SciTech Pu. (29) 39. Anlog Devices, MT-94: Microstrip nd Stripline Design - Anlog Devices, [Online] Aville t: ccessed on July the 7th, M. S. Rni, nd H. Ghfouri Shirz, Improvement of Microstrip Ptch Antenn Gin nd Bndwidth t 6GHz nd X Bnds for Wireless Appliction, IET Microwves, Antenns & Propgtion, DOI: 1.149/ietmp , C. Gu, C. Li, J. Lin, J. Long, J. Hungfu nd L Rn, Instrument-sed noncontct Doppler rdr vitl sign detection system using heterodyne digitl qudrture demodultion rchitecture, IEEE Trnsctions on Instrumenttion nd Mesurement 9, no. 6 (21): D. Oeid, S. Sdek, G. Zhri, nd G. E. Zein, Multitunle microwve system for touchless hertet detection nd hert rte vriility extrction, Microw. Opt. Technol. Lett., 2: doi: 1.12/mop.24877, (21) 43. M. Vlipour, Optimiztion of neurl networks for precipittion nlysis in humid region to detect drought nd wet yer lrms. Meteorologicl Applictions, 23(1), 216, pp M. Vlipour, M. E. Bnihi nd S. M. R. Behhni, Comprison of the ARMA, ARIMA, nd the utoregressive rtificil neurl network models in forecsting the monthly inflow of Dez dm reservoir. Journl of hydrology,476, 213, pp M. Vlipour, M. A. G. Sefidkouhi nd S. Eslmin, Surfce irrigtion simultion models: review. Interntionl Journl of Hydrology Science nd Technology, (1), 21, pp M. Vlipour, Sprinkle nd trickle irrigtion system design using tpered pipes for pressure loss djusting. Journl of Agriculturl Science,4(12), 212, p NHS. Online t: ccessed on July the 7th, 216