1

Software Development for Acquisition and DataManagement in Optical Sensor Networks

A. D. F. Santos, M. F. M. Silva, C. S. Sales, C. S. Fernandes, M. J. Sousa, J. C. W. A. Costa and R. M. Souza.

Abstract—Due to their unique characteristics, optical sensornetworks have found application in many fields, such as in Civiland Geotechnical Engineering, Aerospace, Energy and Oil &Gas Industries. Monitoring solutions based on this technologyhave proven particularly cost effective and can be appliedto large scale structures where hundreds of sensors must bedeployed for long term measurement of different mechanicaland physical parameters. Sensors based on Fiber Bragg gratings(FBG) are the most common solution used in Structural HealthMonitoring (SHM) and the measurements are performed byspecial instruments known as optical interrogators. Acquisitionrates increasingly higher have been possible using the latestoptical interrogators, which gives rise to a large volume of datawhose manipulation, storage, management and visualization candemand special software applications. This work presents tworeal-time software applications developed for these purposes:Interrogator Abstraction (InterAB) and Web-based System(WbS). The innovations in this work include the integration,synchronization, independence, security, processing and real-timevisualization, data persistence and flexibility provided by jointwork of the applications developed. The results showed the useof these softwares in the laboratory and real environments inaccordance with the features proposed.

Index Terms— Software Development, Optical SensorNetworks, Data Acquisition, Data management.

I. INTRODUTION

Currently there has been a significant increase in the use ofsoftware systems in sensor networks [1], [2]. These systemsare made to gather data from sensors in order to buildthe information that defines the real state of the monitoringstructure in function of the time. The extraction of valuableinformation from monitored structures has become a veryimportant and challenging task, mainly due to demand forseveral techniques to compress and condense data as well asto diagnose the state of what is being monitored [3].

In a wireless sensor network with many sensor nodesmeasuring different quantities, as well as an optical sensornetwork using multiplexing, the potential to generate a largevolume of data that can become computationally intractable isquite plausible. In [4] was proposed a wireless network sensorinformation and identification system (WiNS Id) databasewhich archives the data reported by distributed sensors,as well as the implemented support for queries and datapresentation. The system features include real-time supportfor data presentation and visual presentation of sensor nodesreporting in a geographical and temporal context. Additionally,

Applied Electromagnetics Laboratory, Federal University ofPara, Belem, Para, Brazil. E-mails: adam.santos@itec.ufpa.br and{cssj, cindystella, marcojsousa, jweyl}@ufpa.br. This work was supported inpart by CNPq, CAPES and Vale S.A..

WiNS Id provides support for pattern identification and datamining in sensor systems.

A datawarehouse for management of data from wirelesssensors to monitor the habitat of bees was shown in [5].This work proposes a model to extract, transform andnormalize data from sensor networks and load them into adatawarehouse. Thus, this data can be easily analyzed byexperts to assist in the process of decision making. However,this work did not present techniques for compressing datagenerated by sensors, which can be large in volume andimpair the process of information retrieval. The assumptionthat the databases to be integrated are already consolidatedwas considered, but in practice the treatment of large volumesof data ends up being the most costly part of the system.

The SHM process on huge structures, such as dams andbridges, can demand hundreds of sensors. Smart structuresalso can be a demanding application in terms of datarate or samples per second. Optical sensors are beingdeployed in these applications because they are inert,durable, naturally embedded in a reliable optical network andinvulnerable to electromagnetic interference [6]. Differentlyfrom current wireless sensors based on digital systems,these sensors are analogic. They are based on the principlethat the measured information (e.g., temperature, strain,acceleration) is wavelength-encoded in the Bragg reflectionof the grating [7], [8]. Measurand changes are codedon wavelength shift of a given Bragg sensor, which areprocessed by optical interrogator. Interrogators in fiber gratingsensor systems are the measurand-reading units that extractmeasurand information from the optical signals coming fromthe sensor heads. The interrogators usually measure theBragg wavelength shifts and convert the results to measuranddata [9], [10]. This data can be stored in a file or madeavailable to client software that establishes communicationswith the interrogators in accordance with standard protocols.

Although there are already consolidated commercialinterrogators, the data acquisition systems offered with themusually are based on private data formats or databasesincompatible with other systems or brands. Until current date,there are not a commercial or open datawarehouse for SHMcapable of gather data from different technologies and opticalinterrogators. This paper proposes such a monitoring system,composed of two softwares, capable to collect, process,persist, retrieve, manage, and present data from optical sensornetworks. The InterAB System and WbS will be describedbased on their features and their individual contributions to theoverall integrated monitoring system. The paper is organizedas follows: the SHM background is presented in Section II;the InterAB and WbS features are detailed in Sections III

2

and IV, respectively; the Section V demonstrates how theintegration is achieved between softwares; the Section VIpresents and discusses test results; final remarks and futurework are presented in Section VII.

II. THE CORE ACTIVITIES OF THE SHM PROCESS.

The SHM process consists of permanent, continuous,periodic or periodically continuous recording of parametersthat reflect the performance of the structure [11]. Dependingon the type of the structure, its condition and particularrequirements related to a monitoring project, SHM can beperformed in the short term (typically up to few days), midterm (few days to few weeks), long term (few months to fewyears) or during the whole lifespan of the structure.

The representative parameters selected to be monitoreddepend on several factors, such as the type and the purpose of astructure, expected loads, construction material, environmentalconditions and expected degradation phenomena [12]. Ingeneral, they can be mechanical, physical or chemical.The most frequently monitored parameters are presented inTable I. Our work focuses mainly on monitoring mechanicalparameters (strain and acceleration) and partially on physicalparameters (temperature) using optical sensors.

TABLE ITHE PARAMETERS MOST FREQUENTLY MONITORED.

Class ParametersMechanical Strain, acceleration, deformation, displacement

Physical Temperature, humidity, pore pressureChemical Chloride penetration, pH, steel oxidation

The core activities of the structural monitoring process are:selection of monitoring strategy, installation of monitoringsystem, maintenance of monitoring system, data managementand closing activities in the case of interruption of monitoring[12]. In particular, data management can be split in tosub-activities, such as: execution of measurements (readingof sensors), storage of data (local or remote), providing foraccess to data, visualization, export of data, interpretation, dataanalysis and the use of data (warnings and alarms).

The software applications proposed in this work will beresponsible for these sub-activities of data management in thestructural monitoring process.

III. ACQUISITION, PROCESSING AND PERSISTENCE:INTERAB SYSTEM

The InterAB System is a Java application responsible forcommunication, acquisition, compression and data persistence(storage). The InterAB implements the optical interrogator’scommunication protocol. The InterAB operation is shown inFigure 1. The application interacts via TCP/IP socket andthreads with optical interrogators, which collects the samplesgenerated by the sensor network, in order to receive thereal-time monitoring data. The samples are wavelengths shiftsthat indicate changes in the measured parameter by the sensor.The InterAB then processes the data that were transmittedby the interrogator in a format SCPI (Standard Commandsfor Programmable Instruments) and the system communicates

with the database through JDBC connection (Java DatabaseConnectivity) to persist the compressed samples for eachsensor.

Additionally, the data compression techniques implementedin InterAB aim to reduce the amount of records, fromthe structural monitoring with sensor networks, which areinserted into the database [14]. The following techniques wereimplemented: variation in the wavelength and activity.

The compression by variation in the wavelength is suitablefor parameters that varying slowly, such as temperature. Interms of structural monitoring, the temperature usually doesnot show significant changes in a short amount of time, exceptin cases of damage in structure. Therefore, in most cases it isnot necessary to store a large volume of measurement datathat does not represent interesting variations in temperature.The process is performed by comparing the last sample sicollected with the last sample sj that was actually inserted intothe database. If |si − sj | > ∆T , then si is inserted into thedatabase. The parameter ∆T defines the accuracy and intensityof the compression system. When ∆T is large, fewer samplesare inserted into the database. For temperature optical sensorsbased on FBG, the accuracy in Celsius degrees is 0.1oC, thisvalue is a minimal value for ∆T .

The compression technique based on activity is suggestedfor data from acceleration and strain optical sensors. Anessential element of this technique is the identification ofa relevant event or activity. This work proposes the use ofthe absolute difference between the acceleration (or strain)recorded by the last sample and the average of last n samples:

∆si =

∣∣∣∣∣|si| −i−1∑

k=i−n

|sk|

∣∣∣∣∣ , (1)

where ∆si is the absolute difference and si is the last samplecollected by interrogator. The subscript indicates the samplecollection order.

Fig. 1. InterAB System performing data acquisition from multiple opticalinterrogators.

Since the data are stored in the database any desktop or Webapplication can retrieve this information for various purposes,such as visualization, decision making and diagnosis.

3

IV. SECURITY, REAL-TIME VISUALIZATION ANDANALYSIS: WEB-BASED SYSTEM

The Wbs is a Java Web application responsible for retrieve,export, visualization and generation of alerts and alarms frommonitoring data stored into the database by InterAB or anyother application. The WbS operation is shown in Figure 2.First, a user of application performs a request for content (forexample, data export, visualization, interpretation and analysis,or warnings and alarms). Then, the WBS application receivesand processes the request according to what was requested bythe user. Subsequently, based on the user request, the Wbsretrieves the corresponding data in a relational database viaJDBC communication. Finally, the retrieved data is presentedto the user in the form of charts, tables, graphs, files, forms,alerts and alarms.

Fig. 2. The operation of the WbS.

For the monitoring of a target structure was developed aWeb-based application totally independent of the acquisition,processing and persistence performed by InterAB, thereforeonly the information present in the database is required. TheWbS development on the Java programming language togetherwith the JSF (JavaServer Faces) [15] front-end frameworkinheriting the flexibility, integration, mobility and securityafforded by these platforms, once the application becomesvisible for any computer in the Internet.

The WbS is composed of a security mechanismimplemented by Spring Security Web framework that performsauthentication checks and login, thus the most critical systemareas are restricted to people with certain privilege types.Once the communication is established, the user can edit theoptical sensor network settings and procedes tasks such as dataanalysis, export, visualization and the use of data (warningsand alarms). The PrimeFaces [16] Web framework coupled toWeb application aims to implement statistical graphs, tables,forms and computational analysis.

Through virtual version of the optical sensor networkpersisted in the database by InterAB, the WbS creates aWeb version of the sensor network physical topology, whichis an interactive and automatically updatable graphs thatdescribes the current optical sensor network architecture. Eachsensor type has a specific conversion formula, which is called

calibration equation, that converts the peak wavelength ofthe FBG sensor to its corresponding measurement. Theseconversions are performed using first or second orderequations that are provided by datasheets.

The WbS and all features mentioned above are generallyexemplified in the following features diagram shown inthe Figure 3. The diagram illustrates the four layers inthe WbS: visualization layer for user interaction, processinglayer and business rule layer to perform the treatment anddata management, and communication layer (Hibernate) formanipulation and data persistence in the database.

Interface Processing

Try login

Authentication

Monitoring results

Show results

System management

Sucess/failure

Request query

Request query

Business Rule

Select query

Data

Select query

Data

Insert/Delete query

Answer

DAO Hibernate

Database

Request query

Fig. 3. WbS features diagram.

V. FLEXIBILITY, INDEPENDENCE AND INTEGRATIONBETWEEN INTERAB AND WEB-BASED SYSTEM

The interAB and WbS are applications that can workindependently or jointly. The interaction between softwaresare possible by sharing the same database, as shown in Figure4. The InterAB System performs the acquisition, processingand data persistence, at the same time populating a updatetable that records data inserts in the database. These updatesare often checked by the WbS in order to update the databeing made real-time available to the user. In the applicationscontext, Hibernate [13] is characterized as a communicationinterface between the database and the systems, since allpersistence operations and data reading is actually performedby the queries generated by this framework that maps objectsautomatically for the relational database.

InterAB and WbS can ensure high flexibility in the opticalsensor networks monitoring. This flexibility is made feasibledue to the fact that these applications easily adapt to themost types of interrogators and sensors. For interrogators isnecessary to know its communication protocol and samplingrate. And for optical sensors must be known its referencewavelength and calibration equation.

VI. RESULTS

Aiming to present the main features of the InterABSystem and WbS, two tests were proposed in differentenvironments. The first environment is a laboratory test,where temperature is measured under controlled conditions.

4

Interrogator

InterAB

Database

Sensor

TCP/IP

JDBC

Web ApplicationLaptop

JDBC

Fig. 4. Integration between InterAB and Web-based System.

The second environmentis is a real test performed on metalwalkway with approximately 40 m long. The measurementscollected were temperature, strain and acceleration from 8sensors installed along the metal walkway. Each environmentused different interrogator equipments and set of sensors, asshown in Table II.

TABLE IICOMPONENTS FOR EACH ENVIRONMENT

Laboratory testInterrogator Sensor

Type Rate Type #

Rack-Mountable BraggMETER 1 S/sTemperature 3Strain 1Acceleration 0

Real testInterrogator Sensor

Type Rate Type #

Industrial BraggMETER 100 S/sTemperatue 3Strain 3Acceleration 2

A. Test in a laboratory

The scheme for the test is presented in Figure 5. Thisscheme presents a temperature sensor in the three channelsof the optical interrogator and a strain sensor in the otherchannel. The test carried out the temperature monitoring bythree sensors (WTS) and the strain monitoring by one sensor(WSS). The sensor network topology recognized by InterABand visualized by WbS is shown in Figure 6.

Fig. 5. Scheme for test in a laboratory.

Fig. 6. Sensor network topology recognized by InterAB.

The collected, processed and data persisted by InterABcould be visualized by WbS in Figure 7. In this test theInterAB System performs filtering on data generated by opticalsensors. A variation in the peak wavelengths of the temperaturesensors occurs due to a variation in the temperature. Thefiltering was performed only on data that were consistent withcertain wavelength variation set up previously. The graphicsshow the real-time variation of the data collected, beingupdated automatically with the last new one hundred samplessaved in database.

0 5 1 0 1 5 2 0 2 5 3 0 3 5

2 1 , 0

2 1 , 5

2 2 , 0

2 2 , 5

2 3 , 0

2 3 , 5

2 4 , 0

2 4 , 5Te

mpera

ture (

ºC)

S a m p l e s

W T S 1 W T S 2 W T S 3

Fig. 7. Visualization provided by WbS.

B. Test in a real environment

The real testing environment is characterized by a metalwalkway with approximately 40 m long. The walkway andexamples of FBG sensors are shown in Figure 8. An opticalcable of approximatelly 500 m connects the sensor networkinstalled in the walkway with the interrogator located in thelaboratory. The structural monitoring is carried out by a opticalsensor network composed of three temperature sensors, threestrain sensors and two acceleation sensors.

This test presents the features export and preview dataprovided by WBS. The screen regarding this functionalityis shown in Figure 9. The user selects the sensor types andchooses the intervals to select the data. Using the export buttonthe data are saved in an export file format recognized byOctave, Matlab, R and Weka. The update button shows a

5

Laboratory

FC/APC conectorProtection

box

Fig. 8. Scheme for test in a real environment.

preview of data that are available for download, as can beseen in Figure 10.

Fig. 9. Export and data preview.

Acc

ele

ratio

n (

g)

0.4

0.6

0.8

1.0

-1.0

-0.8

-0.4

-0.6

Samples

ACC_1

Fig. 10. Preview of accelerometer sensor data.

VII. FINAL REMARKS AND FUTURE WORKS

The paper presented a monitoring system for optical sensornetworks composed of InterAB and WbS which togetherperform the acquisition, processing, persistence, managementand visualization of data obtained from the interrogationsystems. The results obtained during tests in a laboratory andin a real environment demonstrate the efficiency, robustnessand flexibility of the system for different types of sensors,

optical interrogators and environments, ensuring atomicity,consistency, isolation and durability of persisted data byInterAB and displayed by WbS. For the next steps we intend tointegrate the system features such as fault detection, structuraldamage detection, new filtering techniques and develop a Webservice that can perform a more robust data management.

REFERENCES

[1] V. Vaidehi and D. Sharmila Devi. Distributed database management andjoin of multiple data streams in wireless sensor network using queryingtechniques. In International Conference on Recent Trends in InformationTechnology (ICRTIT), pages 583–588, 2011.

[2] H. Kawashima. A database infrastructure for supporting applicationsof ubiquitous sensor networks. In 5th International Conference onNetworked Sensing Systems (INSS), pages 252–252, 2008.

[3] Jiang Dong, Tian Mao, Liu Wen-Chao, and Pan Yong-Cai. Design andimplementation of a novel wireless sensor network system terminalbased on embedded web server and database. In InternationalConference on Automatic Control and Artificial Intelligence (ACAI),pages 772–775, 2012.

[4] P. Morreale and R. Suleski. System design and analysis of aweb-based application for sensor network data integration and real-timepresentation. In 3rd IEEE Systems Conference, pages 201–204, 2009.

[5] R. A. Costa and C. E. Cugnasca. Use of Data Warehouse to Manage Datafrom Wireless Sensors Networks That Monitor Pollinators. In EleventhInternational Conference on Mobile Data Management (MDM), pages402–406, 2010.

[6] Jose Miguel Lopez-Higuera, Luis Rodriguez Cobo, Antonio QuintelaIncera, and Adolfo Cobo. Fiber Optic Sensors in Structural HealthMonitoring. Journal of Lightwave Technology, 29(4):587–608, February2011.

[7] D. Tosi, M. Olivero, G. Perrone, A. Vallan, and L. Arcudi. Simple fiberBragg grating sensing systems for structural health monitoring. In IEEEWorkshop on Environmental, Energy, and Structural Monitoring Systems(EESMS), pages 80–86, 2009.

[8] P. Antunes, H. Lima, H. Varum, and P. Andre. Static and dynamicstructural monitoring based on optical fiber sensors. In 12thInternational Conference on Transparent Optical Networks (ICTON),pages 1–4, 2010.

[9] N. Yang. Technologies for structural test and monitoring: The modernapproach. In IEEE AUTOTESTCON, pages 1–5, 2010.

[10] R. Adhami. Autonomous structural monitoring using fiber bragg grating.In International Conference on Computer Systems and IndustrialInformatics (ICCSII), pages 1–4, 2012.

[11] Charles R. Farrar and Keith Worden. Structural health monitoring : amachine learning perspective. John Wiley & Sons, Ltd, 2013.

[12] Branko Glisic and Daniele Inaudi. Fibre Optic Methods for StructuralHealth Monitoring. John Wiley & Sons, Ltd, 2007.

[13] JBoss Community. Hibernate ORM. http://www.hibernate.org/, Nov,2013.

[14] A. D. F. Santos, M. J. Sousa, C. S. Sales, M. F. M. Silva, C. S. Fernandes,and J. C. W. A. Costa. Filtering Techniques for Data Persistence in FBGOptical Sensor Networks. In Simposio Brasileiro de Banco de Dados(SBBD), pages 31–36, 2013.

[15] Project Kenai. JavaServer Faces technology. https://javaserverfaces.java.net/, Nov, 2013.

[16] PrimeFaces. PrimeFaces 4.0 - Ultimate JSF Component Suite. http://primefaces.org/, Nov, 2013.