Rb

Ya

b

a

ARRAA

KHOMRBP

1

eohecoiactfppr

0d

Sensors and Actuators A 172 (2011) 359– 364

Contents lists available at ScienceDirect

Sensors and Actuators A: Physical

jo u rn al hom epage: www.elsev ier .com/ locate /sna

eal-time measurement of multipoint hetero-core fiber optic binary sensorsased on optical pulse loss change

usuke Araia,∗, Michiko Nishiyamab, Kazuhiro Watanabea

Information Systems Science and Engineering, Soka University, JapanAerospace Research and Development Directorate, Japan Aerospace Exploration Agency, Japan

r t i c l e i n f o

rticle history:eceived 28 November 2009eceived in revised form 28 March 2011ccepted 6 June 2011vailable online 24 June 2011

eywords:etero-coreptical fiber sensorultipoint

eal-time measurementinary sensor

a b s t r a c t

Hetero-core fiber optic sensors can transmit sensing and communication signals on a single fiber optictransmission line and have numerous advantages for environmental information monitoring such ashome security. Moreover, these sensors are cost effective due to their temperature independence andlight-intensity-based measurements. We have previously developed a hetero-core fiber optic binarysensor that can be connected in series to detect the number of doors and windows that are openedor closed. In this paper, we propose an improved method for using hetero-core fiber optic binary sensorsthat are connected in series, which are referred to as binary switches. A unique pulse loss change enablesthe states of the connected switches to be identified. As a result, the total optical loss in the transmissionline is reduced. Therefore, the number of binary switches connected in series can be increased on asingle transmission line. The unique pulse loss peaks can be controlled by the action of a flat springand by adjusting the position of the flat spring inside the binary switch module. Typical pulse peaks of

ulse loss change each binary switch are from 0.13 to 0.75 dB in the positive direction and from −0.47 to −0.03 dB in thenegative direction, while the typical insertion loss is from 2.23 to 2.61 dB, depending on the positionof the hetero-core segment within the binary switch module. The connection of two binary switches inseries is successfully demonstrated for monitoring the optical loss change on a single transmission line.The results of the present study show that the number of binary switches connected in series can beincreased significantly on a single transmission line.

. Introduction

Recently, the demand for security and monitoring systems forlderly individuals who require care has increased as the numberf crimes involving private homes and elderly people living aloneas increased. These systems can obtain information on the pres-nce of individuals, for example, by monitoring the opening andlosing of doors and windows to track the entering and leavingf rooms in the intelligent environment. Conventional monitor-ng systems that employ electromagnetic devices include camerasnd proximity sensors. However, cameras can evoke feelings of dis-omfort from those being monitored, and proximity sensors tendo be adversely affected by electromagnetic interference. There-ore, such conventional monitoring systems are problematic forrivate spaces, hospitals and nursing homes. In such cases, multi-

oint sensors without cameras that have electromagnetic inductionesistance are desirable alternatives.

∗ Corresponding author.E-mail address: arabou 34@soka.gr.jp (Y. Arai).

924-4247/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.sna.2011.06.004

© 2011 Elsevier B.V. All rights reserved.

Fiber optic sensors have several advantageous characteristics,such as electromagnetic induction resistance and ease of use underadverse environmental conditions. In addition, the sensors areflexible and light weight. Fiber optic sensors have been adoptedas structural health monitoring systems [1,2]; in particular, fiberBragg grating (FBG) sensors [1–5] have been extensively devel-oped to detect minute strain by measuring wavelength shifts witha diffraction grating in the core of an optical fiber. FBG sensors havethe ability to be connected in series on a single transmission lineusing wavelength division multiplexing [4]. However, FBG sensorshave shortcomings such as temperature dependence [5]. As a result,their cost effectiveness is limited by the need to compensate fortemperature fluctuation.

Multipoint sensing schemes based on optical intensity mea-surements with an optical time domain reflectometer (OTDR) orBrillouin optical time domain reflectometer (BOTDR) have alsobeen proposed [6,7]. OTDR requires several minutes to determinethe optical loss by integrating the strength of the returning pulses as

a function of time, because OTDR measure weak backward Rayleighscattering. Although BOTDR has better distance resolution thanOTDR (a few centimeters versus a few meters, respectively), BOTDRis affected by the temperature dependence of Brillouin scattering.

3 Actuators A 172 (2011) 359– 364

Cl

ocstafiwAtmowfmsmtstl

tifAtomsm(rtfitlfcsfH

2

2

ohmaidTidfilisat

Fig. 1. Hetero-core fiber optic binary sensor; (a) schematics of sensor module and

60 Y. Arai et al. / Sensors and

onsequently, a new multipoint sensing scheme using transmittedight is required.

In contrast of conventional fiber sensors, the hetero-core fiberptic sensors [8–11] that we have developed, which measure opti-al loss caused by bending of their hetero-core segments, haveeveral advantages. The first is temperature independence [8], andhe second is cost effectiveness due to their simple fabricationnd light-intensity-based measurements. Using this hetero-coreber optic sensor, we have developed a strain sensor [9] and aearable motion capture sensor [10] for use in smart structures.

multipoint sensing scheme with hetero-core fiber optic sensorshat use OTDR has also been proposed [9]. However, to conduct

ultipoint measurements in real time on the order of a few tensf hertz, a sensing scheme that measures transmitted intensityithout time averaging is necessary. Focusing on the time dif-

erentiation in the optical loss of the hetero-core fiber sensors, aultipoint sensing scheme has been proposed that identifies the

ensors connected on a single fiber transmission line [11]. In thisethod, however, the optical loss from the working sensors on the

ransmission line accumulates, and thus the number of the sen-ors connected in series on a single transmission line is limitedo the detectable range of the measuring instruments for opticaloss.

In this report, we propose an improved method for measuringhe pulse loss change of hetero-core fiber optic sensors connectedn series, which restricts the sensor information to binary form,or example, whether doors and windows are opened or closed.

unique optical loss peak enables the connected sensors sys-em to identify the states of the sensor without accumulatingptical loss, and as a result, the total optical loss in the trans-ission line is reduced. As a result, this allows the number of

ensor connected in series to be increased on a single trans-ission line. This sensor can detect whether a button is pushed

ON) or released (OFF) by measuring optical loss. Thus, we canegard this binary sensor as an optical binary switch. We testedhe improved method to adjust the sensitivity of the hetero-coreber optic binary sensors in multipoint sensing, and demonstratedwo binary sensors connected in series on a single transmissionine. This work presents the concept that the monitoring systemor security to break-in through the doors and windows in homeould be constructed by means of hetero-core fiber optic sen-ors in series and a single transmission fiber line, which is usedor the internet communication, in real-time such as a few tensz.

. Binary switch module

.1. Hetero-core fiber optic sensor

Fig. 1 shows the internal structure of the hetero-core fiberptic sensor and the newly developed binary switch module. Theetero-core fiber optic switch consists of an optical fiber trans-ission line and an inserted hetero-core segment with a length

s short as a few millimeters that was inserted by fusion splic-ng. The fiber transmission line and the inserted fiber have coreiameters of 9 �m and 5 �m, respectively, as shown in Fig. 1.he length of the inserted fiber used in the present experiments approximately 1–2 mm. The transmitted light is lost as leakageue to the different core sizes. At the spliced interface betweenbers with different core diameters, transmitted beam partially

eaks into the cladding layer. The bending curvature, whose radius

s in the range of several 10 cm to 10 mm, changes the angle ofpliced interface, as a result, induces the light leakage and radi-tion loss at the fiber with the core of 5 �m in diameter dueo the smaller core diameter than the transmission fiber line.

optical fiber; (b) photographs of sensor module and (c) an oblique perspective viewof the arrangement of the flat spring, crew and fiber line.

In our previous work, it was found that the optical loss of thehetero-core fiber optic sensing element increases monotonically

as the hetero-core segment was bent. Inserting the hetero-coresegment into the transmission line decreases the overall signalby less than 1 dB when unbent, which is an extremely smallloss.

Y. Arai et al. / Sensors and Actuators A 172 (2011) 359– 364 361

ro-cor

2

fibwatsbFoatsthgicoiNtisotilittts

Fatis

Fig. 2. Experimental set-up for hete

.2. Binary switch module

The improved binary switch module consists of a hetero-coreber optic sensor affixed at points A and B, a flat spring, a pushutton, and two fiber guides as shown in Fig. 1. The flat spring,hich is composed of stainless steel, is clamped by a clamping box

nd set into the switch module. When the push button is pressed,he flat spring gradually induces slight curvature of the hetero-coreegment as indicated by the dashed line in Fig. 1(a). The photos ofinary switch modules are shown in Fig. 1(b) and (c). As shown inig. 1(b), the flat spring touched to the fiber line. Fig. 1(c) shows anblique perspective view of the arrangement of the flat spring, crewnd fiber line. The crew can push the flat spring without touching tohe fiber line, because the height of the crew is lower than the flatpring as shown in Fig. 1(c). Pressing the button completely flat-ens the spring and frees the claw, thus immediately returning theetero-core to its initial position by means that the fiber line couldo through the bottom side of the crew. When the push buttons pressed, the binary switch module induces a positive pulse losshange. When the button is released, the flat spring is bent in thepposite direction, and the curvature of the hetero-core segmentncreases compared with the curvature when the button is pressed.ext, the flat spring immediately returns to its initial position, and

he binary switch module induces a negative pulse loss change. Thismproved system of switches connected in series can detect thetate of all switches by focusing on the change in the total pulse lossn the transmission fiber line rather than the total loss. It is essen-ial that the pulse loss change of each binary switch be adjustablen order to increase the number of switches on the transmissionine. Using the hetero-core fiber sensing technique, the sensitiv-ty of the optical loss change can be easily adjusted by changinghe length of the inserted hetero-core segment and the position ofhe hetero-core segment in the binary switch module. Therefore,he proposed method is suitable for multipoint fiber sensing withensors connected in series.

The protruding length of the flat spring is approximately 7 mm.lat springs in the switching module with thickness of 0.10, 0.07

nd 0.04 mm were used in the present study. The distance betweenhe base of the flat spring and the claw on the push button, whichs defined as d-mm, can be controlled by an adjustment screw ashown in Fig. 1(b). This mechanism allows the contiguous region

e fiber optic binary switch module.

of the flat spring and the claw to be adjusted in order to alter theloss change peak induced by the switch module. The fiber lengthbetween the fixed point A and the center of the hetero-core seg-ment is defined as X-mm. The external dimensions of the moduleare 55 mm × 51 mm × 8 mm, which is sufficiently large, thus allow-ing the unit to be mounted on a fixed stage.

3. Experimental set-up

Fig. 2 shows the experimental set-up for a hetero-core fiberoptic binary switch. A light-emitting diode (LED) that emits a cen-tral wavelength of 1.31 �m with half bandwidth of several tens nmis used as the light source for the binary switch. The propagationlight power was approximately −38 dBm due that the propagationbeam needed to be coupled into the single-mode fiber whose corediameter is as small as 5 �m. The binary switch is equipped with ahetero-core fiber sensor with an inserted fiber with lengths of 1.0or 2.0 mm. The optical loss change is measured with a photo diode(PD) and the data is acquired through the A/D converter, which havethe resolution of 12-bits with 10 kS/s. in sampling rate, in real timewith a laptop computer. The binary switch is mounted on a fixedstage and its button is pressed by a movable stage. The movablestage is driven by a stepping motor operated by a stage controllerat a precision of 200 steps/deg. The stage controller is connectedvia a general purpose interface bus (GP-IB) to the laptop computer.

4. Results and discussion

4.1. Real-time response of hetero-core fiber optic binary switch

Fig. 3(a) and (b) show the real-time response of the hetero-corefiber optic binary switch equipped with inserted hetero-core seg-ments of 1.0 and 2.0 mm, respectively, to the action of the buttonbeing pressed and released in turn. The initial insertion loss of thebinary switches with hetero-core segments of 1.0 and 2.0 mm weregreater than 1 dB, reaching values of 1.50 and 2.41 dB, respectively,

because the initial bending loss of the hetero-core fiber optic sen-sor was induced when the fiber sensor was placed in the switch.Pulse loss change in the positive direction was induced by press-ing the button, and oppositely, the pulse loss change in the negative

362 Y. Arai et al. / Sensors and Actuators A 172 (2011) 359– 364

Ffi

dscttItbbtfimtaOatm2wotfif

4p

csv

Positive pulse peak

Negative pulse peak

The distance from the base of the flat spring to the claw on thepush button, d-mm, is varied to adjust the optical loss pulse peaks.Fig. 6(a) and (b) show the pulse peaks in the positive and negativedirection, respectively, of a hetero-core fiber optic binary sensor

ig. 3. Real-time response of hetero-core fiber optic binary switch with insertedber lengths of (a) 1.0 and (b) 2.0 mm.

irection was induced by releasing the button, as shown in Fig. 3. Ashown in Fig. 3(a), when the button was pushed gradually, the opti-al loss monotonically increased to approximately 0.39 dB. Whenhe button pressing action was complete, the optical loss returnedo 0 dB due to the return of the flat spring to its initial position.n the same way, as the button was released, the loss decreasedo approximately −0.34 dB and then returned to 0 dB when theutton had been completely released. If the button was releasedefore being completely pressed, the flat spring came free fromhe claw. As a result, the switch exhibited a value that differedrom the unique loss peak value. However, for an incomplete press-ng action, the speed at which the optical loss returned to 0 dB

ight be slower than that for a complete pressing action; thus,he proposed detection scheme based on the optical loss peak ispplicable for determining only the switch’s state, namely, ON orFF. As shown in Fig. 3(b), the positive and negative loss peaks werepproximately 1.09 and −0.89 dB, respectively. The pulse peak inhe 2.0-mm hetero-core segment was larger than that in the 1.0-

m segment (Fig. 3), since the hetero-core fiber sensor with a-mm inserted segment was more sensitive to bending. It is note-orthy that the optical loss was maintained at 0 dB, except when

ptical loss changed due to action of the button. This result indicateshat multiple sensors can be deployed along a single transmissionber line, since the total optical loss of a single fiber line is derived

rom only the insertion loss of the hetero-core binary switches.

.2. Relation between pulse peak and hetero-core segmentosition

Fig. 4 shows the variation in optical loss peaks for the hetero-ore binary switch as a function of the position of the hetero-coreegment for 20 trials. As shown in this figure, the optical loss peaksaried from approximately 0.21 to 0.61 dB in the positive direction

Fig. 4. Variation of optical loss peak for hetero-core fiber optic binary switch asfunction of position of hetero-core segment.

and from −0.42 to −0.14 dB in the negative direction, depending onthe position of the hetero-core. The fiber length between the fixedpoints A and B was fixed at 65 mm; the hetero-core segment waslocated between these fixed points. The position of the hetero-corealong the fiber length, X-mm as shown in Fig. 1(b), was shifted insteps of 2 mm from X = 20 up to 32 mm. The maximum sensitivitywas observed when the hetero-core segment was at X = 24 mm forpulse loss in both the positive and negative directions. The positionof the hetero-core segment that induced the largest pulse peak wasslightly to the left of the hetero-core segment at X = 26 mm shownin Fig. 1. This is because the shape of the fiber is asymmetric inthe switch module when the flat spring alters the curvature of thehetero-core segment.

Fig. 5 shows the relationship between the positive and negativeoptical loss peaks and the insertion loss of the hetero-core fiberoptic binary sensor for various positions of the hetero-core seg-ment, X-mm. The positive loss peak and the insertion loss movedin synchrony, while the negative loss peak decreased with increas-ing insertion loss. The insertion loss changed with the hetero-coreposition, X-mm, in the range from 2.32 to 2.61 dB. This is becausethe initial curvature of the hetero-core segment depends on X-mm.This result indicates that a trade-off exists between the positiveand negative pulse peaks, which is related to the position of thehetero-core segment, which governs the insertion loss.

4.3. Method of adjusting optical loss peaks

Fig. 5. Relationship between optical loss peaks, which are positive and negative,and insertion loss of heterocore fiber optic binary switch.

Y. Arai et al. / Sensors and Actuators A 172 (2011) 359– 364 363

a

b

Fig. 6. Characteristics of hetero-core fiber optic binary switch for (a) positive and(

fnip0apffltatfltmweddfltasaaptibm

Switch A

Switch B

ON

OFF

ON

OFF

Threshold level

Threshold level

b) negative loss peaks at various distances, d, over 20 trials.

or various distances, d, and three values of the flat spring thick-ess. The distance, d, was adjusted in steps of 0.025 mm. As shown

n Fig. 6(a), flat springs with thickness of 0.04, 0.07 and 0.10 mmroduced positive pulse peaks ranging from 0.38 to 0.87 dB, from.13 to 0.75 dB and from 0.20 to 0.64 dB, respectively. Similarly,s shown in Fig. 6(b), these flat springs produced negative pulseeaks ranging from −0.30 to −0.03 dB, from −0.47 to −0.03 dB androm −0.47 to −0.16 dB, respectively. Due to its low thickness, theat spring with thickness of 0.04 mm was not straight in the ini-ial state. Therefore, the motion of the flat spring was enlargednd the optical loss peaks increased in the positive direction. Onhe other hand, the pulse peaks in the negative direction for thisat spring were smaller than those of the other flat springs due tohe relatively small motion when the button was released. Further-

ore, the insertion loss for the flat spring with thickness of 0.04 mmas lower than that of the others for the same reason. The switch

quipped with the flat spring with a thickness of 0.04 mm, whoseistance d was less than 6.23 mm, could not be completely flippedue to its softness. On the other hand, the switch equipped with theat spring with thickness of 0.10 mm also could not be flipped whenhe distance, d, was greater than 6.40 mm due to its hardness. Thedjustable range of 0.04 mm to 0.10 mm for the thickness of the flatpring was relatively narrow due to the hardness of the springs. As

result, the flat spring of 0.07 mm in thickness was most useful fordjusting the pulse peak because of the broad dynamic range of itsulse peak. In these experimental results (Fig. 6), it was clear thathe optical loss peak of the binary switch could be adjusted eas-ly by tuning the adjustment screw and that the newly proposed

inary switch could be given a unique pulse peak for connection ofultiple fiber optic binary switch sensors in series.

Fig. 7. Real-time response in optical loss change of two tandem binary switches.

4.4. Identification of two binary switches connected in series

Fig. 7 shows the real-time response of the optical loss changefor two binary switches A and B connected in series. The binaryswitches A and B have unique loss peaks due to the insertedhetero-core segments having differing lengths of 1.0 and 2.0 mm,respectively. As shown in Fig. 7, the loss peaks of switches A andB have unique values in both the positive and negative directions.The unique loss values of switch A were approximately 0.39 dB inthe positive direction and −0.35 dB in the negative direction. Simi-larly, switch B had unique loss peaks of approximately 1.09 dB and-0.84 dB. This result indicates that the states of the switches couldbe successfully identified by detecting the change in the pulse losspeak. For example, in order to determine which switch changedto the ON state, the threshold level of the optical loss change wasset at 0.6 dB, as shown in Fig. 7. In the same way, determination ofwhich switch was in the OFF state was accomplished by setting athreshold level of −0.6 dB. As expected, the total optical loss on thesingle transmission line did not accumulate due to the state of theswitch. The total loss on the fiber transmission line was approxi-mately 3.94 dB, which was the sum of the insertion loss of switchesA and B, and the connection loss of the SC connecter. Therefore,the method of adjusting the optical loss peaks, as shown in Fig. 6,allowed the number of the binary switch modules connected inseries to be increased, and the number of switches on a single trans-mission line was primarily dependant on the insertion loss of thebinary switches.

Fig. 8 shows the simulation results for two overlapping opticalpulses of switches A and B. Overlapping optical pulses complicateidentification of the working switch on a single transmission linebecause optical loss accumulates. Using the time differential value,however, two peaks of overlapping pulses can be correctly distin-guished, as shown in Fig. 8(b). The time differential method candetect the return of the flat spring because the speed at which thepulse peak returns to 0 dB is sufficiently faster than the speed atwhich the button is pressed. Therefore, the difference in opticalloss before and after the return of the flat spring to its initial posi-tion possibly allows the unique loss peaks of the switches to bedistinguished. As shown in Fig. 8(a), the total optical loss changeof switches A and B connected in series was 1.48 dB in the posi-tive direction when the buttons of each were pressed in unison.Consequently, the switches A and B were difficult to identify. How-

ever, focusing on the optical loss change during the return of theflat spring, as shown in Fig. 8(a), the loss change of switches A andB connected in series in the positive direction was 1.09 dB, which

364 Y. Arai et al. / Sensors and Actuat

Time-differential value of switch A+B

Switch A

Switch B

Switch A+B

Returning pulse : 1.09 dB

Pulse peak of switch B : 1.09 dB

a

b

FB

iTbfw

5

osctnsegtctbsiptbsmcfi

[

ig. 8. Simulation result of two overlapping optical pulses induced by switch A and (a) optical loss change and (b) time differential values.

s almost the same as the pulse peak of switch B, namely, 1.09 dB.hese results indicate that switches with overlapping pulses cane distinguished, and that this method, whereby pulse loss is dif-erentiated with respect to time, can limit misidentification of theorking switches.

. Conclusions

In this paper, we proposed an improved sensing method basedn pulse loss change for using hetero-core fiber optic binarywitches connected in series. The proposed binary switch systemonnected in series enabled the total loss on the transmission lineo be reduced. As a result, the number of binary switches con-ected in series could be increased. The pulse loss changes wereet by adjustment of a flat spring in the binary switch module tonable the binary switches to be uniquely identified on the sin-le transmission line. The pulse peak varied with the position ofhe hetero-core segment in the binary switch module and wasorrelated with the insertion loss of the binary switch. The posi-ive and negative pulse peaks of the proposed binary switch coulde adjusted by altering the distance between the base of the flatpring and the claw attached to the push button. The flat springnduced bending of the hetero-core segment, thus producing aulse with an adjustable range from 0.13 to 0.75 dB in the posi-ive direction and from −0.47 to −0.03 dB in the negative directiony using a flat spring with thickness of 0.07 mm. The two binary

witches A and B were successfully connected in series to conductonitoring on a single transmission line by measuring optical loss

hange. The proposed method allows the number of hetero-coreber optic switches connected in series on a single transmission

[

ors A 172 (2011) 359– 364

line to be increased. Additionally, the sensitivity of hetero-corefiber optic switch can be increased by means of adjustment to thelength and core-diameter of hetero-core portion and lengtheningthe controllable range of pushing button. The limitation of a num-ber of the proposed hetero-core fiber optic switches is consideredto derive from the detectable sensitivity of optical power detec-tor, threfore, to depend on the accumulation of initial insertion andbending losses of hetero-core fiber. Since the hetero-core fiber sen-sor induces the low initial insertion loss of less than 1 dB, a few tenshetero-core fiber optic switches could be connected in series usingthe proposed sensing method in real time. As a result, the hetero-core fiber optic binary switch system can potentially be employedas an appropriate and cost-effective home-security sensing net-work and has the capability to detect environmental informationon a ubiquitous network.

Acknowledgments

This research was supported in by the “Collaboration with LocalCommunities” Project for Private Universities on “Developmentof ubiquitous-Monitoring-Network Based on Distributed SensorNodes using Local Positioning / Optical Sensory Nerves and theirIndustrial Applications” and by matching funds from the Ministryof Education, Culture, Sports, Science and Technology of Japan(MEXT), 2006-2010.

References

[1] M. Majumder, T.K. Gangopadhyay, A.K. Chakraborty, K. Dasgupta, D.K. Bhat-tacharya, Fibre Bragg gratings in structural health monitoring—present statusand applications , Sens. Actuators A 147 (2008) 150–164.

[2] J. Leng, A. Asundi, Structural health monitoring of smart composite materialsby using EFPI and FBG sensors , Sens. Acutuators A 103 (2003) 330–340.

[3] A.D. Kersey, M.A. Davis, H.J. Patrick, M. LeBlanc, K.P. Koo, C.G. Askins, M.A.Putnam, E.J. Friebele, Fiber grating sensor , J. Lightwave Technol. 15 (1997)1442–1463.

[4] A.D. Kersey, T.A. Berkoff, W.W. Morey, Multiplexed fiber Bragg grating strain-sensor system. with a fiber Fabry–Perot wavelength filter , Opt. Lett. 18 (1993),No. 16.

[5] Y.J. Chiang, L. Wang, H.S. Chen, C.C. Yang, W.F. Liu, Multipoint temperature-independent fiber-Bragg-grating strain-sensing system employing an optical-power-detection scheme , Appl. Opt. 41 (9) (2002) 1661–1667.

[6] C. Li, Y.-M. Zhang, H. Liu, S. Wu, C.-W. Huanga, Distributed fiber-optic bidi-rectional strain–displacement sensor modulated by fiber bending loss , Sens.Actuators A 111 (2004) 236–239.

[7] T. Kurashima, T. Horiguchi, H. Izumita, S. Furukawa, Y. Koyamada, Bril-louin optical-fiber time domain reflectometry , IEICE Trans. 76-B (4) (1993)382–390.

[8] H. Sasaki, M. Nishiyama, E. Kondo, N. Shinomiya, K. Watanabe, Combinedfiber optic network for. communication and measurement using a hetero-corespliced fiber optic sensor, IOP 19th International Conference on Optical FibreSensors (OFS-19), Proc. of SPIE, vol.7004 700449, Perth Australia, April 2008.

[9] H. Sasaki, Y. Kubota, K. Watanabe, Sensitivity property of a hetero-core splicefiber optic displacement sensor , Proc. SPIE 5579 (2004) 136–143.

10] M. Nishiyama, H. Sasaki, K. Watanabe, Performance characteristics of wearable

embedded hetero-core fiber sensor for unconstrained motion analyses , Trans.SICE 34 (12) (2007) 1075–1081.

11] M. Nishiyama, H. Sasaki, K. Watanabe, Optical intensity-based measurement ofmultipoint hetero-core fiber sensors by the method of time-differentiation inoptical loss , IEEE Sens. J. 8 (7) (2008) 1055–1060.