CONTENTS

CHAPTER 1: INTRODUCTION

01

CHAPTER 2: ISO/IEC/IEEE 21451 REFERENCE MODEL

CHAPTER 3: HARDWARE ARCHITECTURE COMPLIANT WITH IEEE 21451

CHAPTER 4: TRANSDUCER SIGNAL SERVICES
CHAPTER 5: EXPERIMENTAL EVALUATION

1.

CHAPTER 1
INTRODUCTION
Smartlightingisabuzzwordthatdefinesheterogeneouslightingtechnologies:High
intensitydischargelamps(HID),fluorescent,solidstateLEDandOLEDluminaires,
composedofnumeroussmartsensorsandactuators,andwiththepossibilityofincorporatinga
widesetofcapabilitiesandconnectivityinterfaces.Themostimportantintelligentfeaturesare
relatedtoenablingadvancedfunctionssuchasadjustablespectralreproduction,colortunable
lighting,andadaptivedimming,combiningenergyefficiencywiththerealneedsofthe
illuminatedspacetakingintoaccountavailablenaturallighting.Withtheriseindigital
connectivityoptions,smartlightingsystemshavedifferentwiredandwirelessinterfacesoriented
toincrementtheconnectivityinsmartgridsystemsandBuildingManagementSystems
(BMS).Thesesystemsallowustocontrolandmonitormodernandheterogeneouselectrical
equipmentssuchasventilationandlightingdevices.Themaininterfacesconceivedforwired
lightingsystemsareDigitalAddressableLightingInterface(DALI),PowerLine
Communications(PLC),DigitalMultiplex(DMX512)andKNXforintelligencebuildings.
ThemostimportantwirelessphysicalinterfacesfordistributedintelligentdevicesareIEEE
802.11a/b/g/n/ac(WiFi),802.15.1(Bluetooth),IEEE802.15.4(ZigBee,6loWPAN,ISA100,
WirelessHART,MiWi)andISO/IEC18000(RFID)compatibleforautomationapplications.
Nowmodernsmartlightingsystemscontainwiredandwirelessinterfaces,andtheycanbe
harmonizedinasetofISO/IEC/IEEE21451standards[9][12].Thisharmonizationcovers
TransducerElectronicDatasheets(TEDS)formats.TEDSaredifferentelectronicdatasheets
storedinanonvolatileflashmemory.Theseformatsdefinenodecapabilitiesandtransducer
channelconfiguration.Inaddition,ISO/IEC/IEEE21451standardcoversstandardsyntactic
messages,andlogicaltransducerfunctionservicestobeimplementedintheNetworkCapable
Applicationmodule(NCAP)and/ortheTransducerInterfaceModule(TIM).
An interoperable smart lighting system that implements the IEEE 802.15.4 physical layer and
ZigBee Light Link (ZLL) , plus ISO/IEC/IEEE 21451 as application layer, has been
implemented in this study as a practical solution for communication between wireless lighting

devices, colour sensors, occupancy detectors and illuminance sensors.

In order to demonstrate the use of Fuzzy Logic theories in smart lighting systems, a predictive
algorithm adjusts the dimming level based on the occupancy level, work schedule and natural
lighting available. A graphical user interface was designed in Matlab to run the fuzzy algorithm.
Finally, all information is encapsulated in standard message structures to interact with the
Representational State Transfer (REST) web service architecture . The web service runs on a
free platform that contains an open API intended to control intelligent devices for the Internet of
Things (IoT).
Capabilities and connectivity interfaces. The most important intelligent features are related to
enabling advanced functions such as adjustable spectral reproduction , color tunable-lighting,
and adaptive dimming , combining energy efficiency with the real needs of the illuminated space
taking into account available natural lighting. With the rise in digital connectivity options, smart
lighting systems have different wired and wireless interfaces oriented to increment the connectivity in smart grid systems and Building Management Systems (BMS) . These systems
allow us to control and monitor modern and heterogeneous electrical equipments such as
ventilation and lighting devices. The main interfaces conceived for wired lighting systems are
Digital Addressable Lighting Interface (DALI), Power Line Communications (PLC), Digital
Multiplex (DMX512) and KNX for intelligence buildings.
The most important wireless physical interfaces for distributed intelligent devices are IEEE
802.11a/b/g/n/ac (WiFi), 802.15.1 (Bluetooth), IEEE 802.15.4 (ZigBee, 6loWPAN, ISA100,
WirelessHART, MiWi) [7] and ISO/IEC 18000 (RFID) compatible for automation applications .

Now modern smart lighting systems contain wired and wireless interfaces, and they can be
harmonized in a set of ISO/IEC/IEEE 21451 standards . This harmonization covers Transducer
Electronic Datasheets (TEDS) formats. TEDS are different electronic data sheets stored in a non
volatile flash memory. These formats define node capabilities and transducer channel
configuration. In addition, ISO/IEC/IEEE 21451 standard covers standard syntactic messages,
and logical transducer function services to be implemented in the Network Capable Application
module (NCAP) and/or the Transducer Interface Module (TIM).More formally, each NCAP

performs the task of coordinator node while the TIM module will be integrated in the luminar-ies
or in different smart sensors. Both can be organized as a Wireless Sensor and Actuator Network
InordertodemonstratetheuseofFuzzyLogictheoriesinsmartlightingsystems,apredictive
algorithmadjuststhedimminglevelbasedontheoccupancylevel,workscheduleandnatural
lightingavailable.AgraphicaluserinterfacewasdesignedinMatlabtorunthefuzzyalgorithm.
Finally,allinformationisencapsulatedinstandardmessagestructurestointeractwiththe
RepresentationalStateTransfer(REST)webservicearchitecture[18].Thewebservicerunsona
freeplatformthatcontainsanopenAPIintendedtocontrolintelligentdevicesfortheInternetof
Things(IoT).
Thepaperisorganizedasfollows:theISO/IEC/IEEE21451referencemodel(NCAPand
WTIM)ispresentedinsectionII.SectionIIIdescribesthehardwarearchitectureofasmart
luminaireandautonomoussensornodes.Next,insectionIV,thealgorithmintendedforsmart
lightingcontrolisexplained.TheRESTwebservicearchitectureandexperimentalresultsare
presentedinsectionV.Finally,insectionVIconclusionsandconsiderationsrelatedtoenergy
savingswhichcanbeachievedthroughtheproposedsolutionarepresented

CHAPTER 2
ISO/IEC/IEE 21451 REFERENCE MODEL
ThefamilyofISO/IEC/IEEE21451(formerlyIEEE1451)[19]aimstoexpandthesyntactic
interoperabilitythroughastandardizedmessageformatimplementedusingheterogeneoussmart
sensorsandactuators.SeveralTransducerInterfaceModules(TIM)workincooperationwiththe
NetworkCapableApplicationmodule(NCAP)operatingasacoordinatornodeentity.The
NCAParchitectureisshowninFigure1

Fig 1. NCAP architecture

EachNCAPentitydefinesagroupoflogicalinterfaceformats.Thefirstsetofnetwork
interfaceservicesisintendedforRepresentationalStateTransfer(REST)webservices.These
formatsaresuggestedforsmartsensorswithInternetconnectivity.Differentlogicalnetwork
protocolsintendedforthisproposalareTCP/UDP,HTTP,WebInterface,SimpleNetwork
ManagementProtocol(SNMP)andeXtensibleMessagingandPresenceProtocol(XMPP).
ThesecondsetoflogicalformatsthatrunsintheNCAPisdefinedascommoncorenetwork
servicesISO/IEC/IEEE214511.Theseservicesincludessubclassesfordiscovering,transducer

dataaccess,TransducerElectronicDataSheet(TEDS)access,eventnotification,andtransducer
managementinordertoaddress,mapandcontrolanyTIMdevice.
ThethirdsetofformatsinISO/IEC/IEEE214510,formerlyIEEE1451.0,definescommon
controlcommandsbetweentheNCAPandtheTIMdevice.Inaddition,physical communication
capabilitiesaredefinedinISO/IEC/IEEE214512standardforsmartsensorsworkinginwired
networkswhiletheISO/IEC/pIEEE214515standard(formerlyIEEE1451.5)isintendedfor
smartsensorsworkinginwirelesssensornetworks(WSN),(WiFi,IEEE802.15.4,Bluetooth)
whileISO/IEC/IEEE21451.7wasproposedforRFIDdevices.
ThesecondpartoftheISO/IEC/IEEE21451referencemodelisrelatedtotheTransducer
InterfaceModule(TIM).SeeFigure2.

Fig 2. TIM architecture

EachTIMisanembeddedphysicaldevicethatimplementsareducedsetoflogicalservices.
ThetypicalTIMhardwarearchitecturecontainsanultralowpowermicroprocessorunit(MCU)

andanonvolatilememorytostorethenodeconfigurationinanelectronicdatasheet(TEDS).
TheIEEE21451referencemodeldefinesdifferentstandardizedTEDSformatscorrespondingto
thetypeof
transducer.TheIEEE214514standarddefinesaphysicaltransducerchannelinamixedmode
interfacewhiletheIEEE214512standarddefineTEDSformatsforserialinterfaces,andboth
standardsareorientedforoperationinwirednetworks.However,theIEEE214517standard
coverssyntacticinteroperabilitybasedonstandardizedTEDSformatsforRFIDdevices.In
addition,theIEEE21451.5standardcoversdifferentstandardizedTEDSformatsfortransducers
(sensorsandactuators)workinginaWirelessSensorNetwork(WSN).

CHAPTER 3
HARDWARE ARCHITECTURE COMPLIANT WITH IEEE 21451
Thesmartlightingsystemisdeployedasanecosystemofsmartsensorslocatedinwalls,
windows,thechilledbeamsystemandtheluminairesinstalledintheceilingoftheoffice.This
sectionisfocusedondescribingtheluminairearchitectureandthesensornodecapabilities,
includinganenergyharvestingmodule

3.1SmartLuminaire

I2C/ADCchannels(illuminancesensor,colourdetection)andadigitalinput/outputpinfor
digitaloccupancysensor

Fig 3. Smart luminaire block diagram and implemented system

TheMCUprocessingunitcontainstheMSP430CPUandthefirmwareimplementsthe
ISO/IEC/IEEE21451TIMreferencemodel.Thedecisiontakingalgorithmcalculatesthe

percentageofdimmingtocontroltheluminaire.OnewirelesstransceiverCC2420isthecoreof
theradiosection.ThisdevicesupportstheIEEE802.15.4physicallayerandtheZigBeeLight
Link(ZLL)applicationlayer.
Thedriveriscomposedofasolidstaterelay.ThissectionusestheCPUI/OportsandDAC
outputandanamplifierforsignalconditioning(110VDC).FinallyacircuitcontainingaBJT
transistorandaprotectiondiodemanagestheload.Thissubsystemisactivateddependingon
occupancydetectionorusagepatternsrelatedtostudiesonenergysavingsinlightingsystems.
Theflashmemorystoresmetadataconfigurationrelatedtothesensorschannelscompliantwith
theIEEE21451.0formatandIEEE21451.5forthephysicalTEDS(IEEE802.15.4plusZigBee
LightLink).
Threedifferentsensorsareimplementedinthesensingsubmodule:illuminance,occupancy
andcolour.ForilluminancestudiestheTSL2563sensorcontainingtwointernalphotodiodes,
wasselectedduetohighdynamicrange(0lx100.000lx)andcapabilitiestodetectphotopic
andscotopicluminosityfunctionstomeasurethevisualperceptionunderdifferentlighting
conditions.Thefirstphotodiodecapturesilluminanceinvisibleandinfraredspectrum(300nm
1100nm)whilethesecondphotodiodecapturesonlytheinfraredrangeoflighting(600nm
1100nm).Theoccupancydetectorisbasedonapassiveinfrared(PIR)digitalIRTECMS
320LPsensor.Itcontainsdualelementsforpyroelectricinfraredsensingandthebodyincludesa
lensandanopticalfilterforpassinginfraredradiation.ForcolormeasurementstheP12347
01CTwasselectedtoadjustforcolorcorrelatedtemperature(CCT).

3.2.AutonomousSmartSensors
For human centric lighting studies several autonomous sensors to measure illuminance were
developed. All of them contain a photovoltaic harvester based on organic solar cells for indoor
operation. EachnodecontainsanenergyharvestingsubmoduleandaTIMunit.Theenergy
harvestingunitcontainsorganicandflexiblesolarcells,adirectcurrent(DC/DC)micropower
regulator,supercapacitorsandanON/OFFswitchtoenablethedevice.ThemainTIM
componentsaretheMSP430CPU,theexternalwirelesstransceiverandtheflashmemoryto

storetheTEDSconfiguration.Thenodeswerelocatedintheofficenearthewindowsandthe
wallstomeasureilluminancevaluesattheheightoftheworkingplane

Fig 4. Block diagram for smart sensor and their placement in windows and office walls

3.3. Firmware Implementation

Fig 5.TIM finite state machine

Figure5representstheoperationalstatesoftheTIMandnodecapabilitiesinaccordancewitha
finitestatemachinemodel.TheautonomousTIMwakesupwhentheharvestercapturesenergy

fromtheDSSCsolarmoduleduetoanincreaseindaylightorartificiallightinglevels.Thefirst

stateistheinitializationstep(init)thatoccursafteraneventrelatedtopoweronor
reset,atthattime,TEDSconfigurationisloadedintotheRAMmemory.The
second
stateistheactivemode.Inthismode,TIMpollseachphysicaltransducerchannel
andisalreadyabletosendorreceiveshortmessages.Thethirdstateisasleepmodethatis
usedtoreduceenergyconsumption.Inthissleepstate,thenodewaitsfortheexpirationofa
timertowakeup.TheNCAPrequestscommandsandtheTIMrespondswithshortmessages
correspondingtothecommandtype.
Aftertheinitializationtakesplace,eachTIMreadstheTEDScalibrationvaluesforthephysical
channels.Oncetheprogramchecksallthevalues,itsavesthemeasureddataintheRAM
memory.Themainlooprunsonceeveryfiveminutes.OnelooprunstheIEEE21451commands
andsortsallarraysusingstandardizedmetadataformats.Afterthesortingproceduretakesplace,
theprogramgoestoactivemodeandstartsthetransceivertosendorreceivemessages.The
meanvaluesofphysicalchannelsareusedbecause,occasionally,itcanhappenthatincorrect
valuesaremeasuredduetopowermanagementrestrictionsrelatedtothenode.Theseincorrect
measuredvalueshavetobedetectedandfilteredfromtruedata.
Inatimeframeof2milliseconds,thelastmeasureddataisstoredinRAMmemoryinorderto
besenttowardstheNCAP.ThecoordinatingNCAPisconnectedthroughaUSBportfroman
industrialPCrunningWindows7operatingsystemandaMatlabapplication.Eachmessagein
theNCAPisencapsulatedinshortstructures,incompliancewiththeIEEE21451standard.
ThesemessagestructuresaresenttotheTIMnodes.WhentheNCAPsendsamessage,itwaitsa
responsefromTIMduring500ms.Ifaresponseisnotreceived,thecoordinatorwaitsfive
minutesbeforetheprocessisrepeated.Then,theIEEE21451packetsaresenttowardsavirtual
serialporttodecodeallmetadatainaJavaScriptObjectNotation(JSON)formatthatrunsina
Matlabgraphicaluserinterfacebeforebeingsentouttothewebservice.

CHAPTER 4
4.1 FuzzyLogicAlgorithm
Lightingcontrolsystemshavebeendefinedhistoricallytominimizepowerconsumption
and/orincreasevisualcomfort.Traditionally,thesestrategiesaredividedintofourdifferent
approaches.Thefirstoneinvolvesupdatingthelightingsystemwithenergyefficientluminaires.
Thesecondapproach,introducesoccupancysensorsinthelightingsystem.Thethird,the
lightingsystemcontainsdaylightharvestingandcontrol.Finallythefourthapproach,alighting
systemfurtherincludesmotorizedblinds.However,inourlightingcontrolsystemprototype,we
attempttocoverthesedifferentapproachesbyconsideringaheuristicmodeltodeterminecontrol
actionsbasedonuserexperienceandwiththesupportofsmartsensorsandactuators.Several
heuristicmodels,consideringartificialintelligencetechniqueshavebeenreportedduringthelast
years.Insomecases,theseheuristicmodelsseektosolveproblemsthattraditionalmethods
cannotworkout.Thesesolutionsareoftenbasedoninnovativestrategies,rulesorprinciplesof
datafusion.Forthisreason,wechoseaFuzzylogicalgorithmtodeterminelightlevelsrequired
ontheworkplaneduringaworkingdayinanoffice.Thealgorithmistreatedasanadditional
IEEE21451serviceandwasdesignedasshowninFigure6.

FIG 6. PIEEE21451 SIGNAL TREATMENT SERVICES

Thissignaltreatmentserviceattemptstoadjustlightinglevelsbysynthesizingasetoflinguistic
controlrulesobtainedfrominformationcollectedduringexperimentalmeasurements.The
computationalalgorithmadjustslightingcomfortlevelsincompliancewithUNE124642012.
TheinputvariablesintheFuzzymodelareoccupancy,daylightlevelontheworkplane,the
positionoftheblinds,thescheduleofworkintheofficeandthedaylightindexeswhichhave
beencalculatedfromsimulationswiththeraytracingsoftwareDaysim.Eachinputvariable
passesthroughaMamdaniFuzzyprocess,usingalogisticsigmoidfunctionandisrepresentedin
afuzzificationprocess.Inthiscase,eachvariableiscodedasafuzzyfunctionandisevaluated
byusingcertainrulesofinferenceasshowninTable1
Table 1. FUZZY LOGIC ALGORITHM

Therulesmustbecombinedinsomemannerinordertomakeadecision.The
algorithmusesinferencerulesthroughbooleanlogicandaggregationmethodsof
thetype(IFORNOTAND)forquantifyingtheinputvariables.Theoutputvalues
mustbedefuzzifiedinordertoresolveasingleoutputvaluefromtheinputdataset.
Thefinaldesiredoutputistheexpectedartificialilluminanceorganizedasa
metadatastandardsyntacticstructure.Table2showsthesyntacticmessagestructure.
TABLE 2 .IEEE21451.0MESSAGESTRUCTURE

TheheadersrelatedtotheNCAPrequestare:ZLLFramecontrol(FC),ZLLdeviceidentifier
(DEVID),ZLLcommandidentifier(CMDID).Inaddition,destinationtransducerchannel,
commandclass(Cmdclass),commandfunction(CmdFunc),lengthoffsetanddatainformation
aretheIEEE21451.0headers.
TheheadersineachTIMresponsetransactionare:ZLLFramecontrol,ZLLdeviceidentifier,
ZLLcommandidentifier.Also,success/failflag,lengthoffsetandmetadatapayloadarethe
IEEE21451.0headers.

Theapplicationlayerwhererunthefuzzylogicalgorithmwasinitiallydesignedas
aMatlabfunction,theNCAPsentacommandtoreadilluminanceandTIM
responsewithasmallmessageof13bytes.Then,theNCAPmergedilluminance

valuesandestimatedoptimizeddimmingvalues.Whentheoutputofthefuzzy
algorithmobtainedaresponse,theNCAPcomparedcurrentdimmingvaluesbefore
sendingarequesttochangethedimmingvalueintheTIMsmartluminaire.A
secondfuzzylogicalgorithmrunsinternallyinaTIMmoteasanadditional
internalIEEE21451service.Inthiscase,thefinalMatlabFuzzylogicFISfileis
convertedtoaCfunctionbyusingtheFIST:MATLABFuzzyInferenceSystemC
Converter[30],andoverallfirmwarecodeisportedinitiallyfromArduinotoward
TinyOS2.1.Thecompiledcodeoccupies220bytesofRAMand21568bytesof
ROMintheTIMnodesbasedontheMSP43016bitmicrocontroller.
Chapter 5
EXPERIMENTALEVALUATION
UserSatisfactionSurvey
Priortheimplementationofthenewlightingsystem,wefocusedonaproperwork
environmentcharacterizationbystudyingtheenergyconsumptionandtheuservisualcomfort.
Differentmeasurementsoftheenergyconsumption,spectralreflectionsurfacesintheroomand
glareanalysiswithaluminancecameraweretaken.Thesedataallowedustodetermine
experimentalvaluesandhelpimprovingtheuservisualcomfortandsensorposition,depending
onthespectralcontributionsofnaturaldaylight.Astheoldlightingsystemlackeddimming
capabilities,thecontributionfactorofartificiallightingwasoptimizedbymeansofaheuristic
fuzzyalgorithm.Usersatisfactionwasassessedbeforeandafterinstallationofthenewlighting
systemthroughasurvey,asshowninfigure10
Table3
FunctiontoreadtheTIMchannel,TEDSandCommands

Table 4
FunctiontoreadtheTIMchannel,TEDSandCommands

Fig 10. Survey impact of the light system

Chapter 6
Conclusion
Inthispaper,asmartlightingsystemthattakesintoaccountuserpreferencesandincludessmart
sensorsandactuatorsbasedonIEEE21451isproposed.AutonomousTIMnodesprovidedwith
energyharvestingtechniquesweredesigned.Areducedsetoffunctionsthatcoverthecoreofthe
newIEEE21451standardwerealsoimplemented.
Thisnewlightingsystemisabletoachieveenergysavingsbetween13%and43%comparedto
theoldfluorescentofficelightinginstallation.Asshowninthepresentedquantitativesurvey
indicators,experimentalresultsshowareductioninoperationalenergycostsandaclearincrease
invisualcomfortfortheoccupantsfrom75%to86.5%.
AnadditionalIEEE21451FuzzylogicservicewasimplementedasaMathlabfunctionto
optimizecontrolstrategiesdirectlyintheNCAP.Inaddition,asecondIEEE21451fuzzyservice
withintheTIMfirmwarecontinuouslyadaptsdimmingvaluestoguaranteeoptimalvisual
comfortatpointofmeasurement.


REFERENCES
1. J.Byun,I.Hong,B.Lee,andS.Park,IntelligenthouseholdLEDlightingsystem
consideringenergyefficiencyandusersatisfaction,
2. R.SrividyaandC.P.Kurian,WhitelightsourcetowardsspectrumtunablelightingA
review,inProc.Int.Conf.Adv.EnergyConvers.Technol.
3. Q.ChunhuiandW.Wei,ALEDcontrolsystemwithbrightnessandcolortemperature
adjustable,inProc.Int.Conf.ControlAutom.Syst.Eng.(CASE)
4. D.Park,Z.Liu,andH.Lee,A40V10W93%efficiencycurrentaccuracyenhanced
dimmableLEDdriverwithadaptivetimingdifferencecompensationforsolidstate
lightingapplications.