DEGREE PROJECT IN TECHNOLOGY,
FIRST CYCLE, 15 CREDITS
STOCKHOLM, SWEDEN 2016
Light Tracking Robot
Navigation using light and colour sensors
MIKAELA KARLÉRUS
BEATA TÖRNEMAN
KTH ROYAL INSTITUTE OF TECHNOLOGY
SCHOOL OF INDUSTRIAL ENGINEERING AND MANAGEMENT
Bachelor Thesis MMKB 2016:41 MDAB102
Light Tracking Robot
Mikaela Karlérus
Beata Törneman
Approved
Examiner
Supervisor
2016-06-07
Martin Edin Grimheden
Baha Alhaj Hasan
ABSTRACT
The increasing demand of making the roads safer has trigged a lot of companies to
developcompleteself-drivingcars.Aself-drivingcarrequiresagreatnumberofdifferent
sensorsasgyros,radars,GPS,tachymetersetc.andadvancedsoftware.Thisthesiswill
focus on the possibilities of using only light sensing devices for a tracking robot and
examinetheadvantagesanddisadvantagesofthis.
Thepurposeistoinvestigatewhichtypeoflightsensorismoresuitableforatracking
robotandwhatthelimitationsofatrackingrobotusingthistechnologyare.
Ademonstratorusingtwolightsensorsforcontrollingspeedanddirectionandacolour
sensortoavoidobstacleswillbebuilt.Apartfromchoosingthemostsuitablesensorfora
light-trackingrobotthesensingdistanceandrangeofthechosenonewillbetested.
To investigate the different light-tracking possibilities and the accuracy of the
demonstrator, the vehicle will be put in an open indoor space with arranged coloured
luminous obstacles. The robot will be tested in both a completely dark room and a lit
room.Theintentionwiththeoutcomeistoseethedifferencesoftherobotsbehaviour
whendisturbancesfromsurroundinglightareaddedasanadditionalaspect.
Theresultsfromthetestarepresentedandtheuseofdifferentsensorsarediscussed.
Thefinalconclusiononusinglightsensingforatrackingrobotisthatitisaneasyand
inexpensivemethod,butitshouldbeusedasacomplementtoothersensingdevicesnot
asastand-alonemethod.
I
II
Kandidatarbete MMKB 2016:41 MDAB102
Light tracking robot
Mikaela Karlérus
Beata Törneman
Godkänt
Examinator
Handledare
2016-06-07
Martin Edin Grimheden
Baha Alhaj Hasan
SAMMANFATTNING
Den ökande efterfrågan på säkra bilvägar har lett till en utveckling av kompletta
självkörande bilar. En självkörande bil kräver ett stort antal olika sensorer som gyros,
GPS,tachymeterosv.samtavanceradprogramvara.Dennaavhandlingkommerfokusera
påmöjligheternaattkonstrueraenspårningsrobotmedbaraljuskännandeenhetersamt
undersökafördelarochnackdelarmeddetta.
Syftetärattavgöravilkentypavljussensorsomlämparsigbästförenspårningsrobotoch
vilkabegränsningarenspårningsrobotsomanvänderdennateknologikommerattha.
Enprototypsomanvändertvåljussensorerförkontrollerandeavhastighetochriktning
samtenfärgsensorförattundvikahinderkommerattkonstrueras.Bortsettfrånattvälja
den mest lämpade sensorn för en spårningsrobot, kommer avståndskänning och
räckningsvidd att testat för den valda sensorn. Roboten kommer att testas i både ett
fullständigtmörktrumochiettupplystrum.
Förattundersökadeolikaljusavkänningsalternativenochnoggrannhetenhosroboten,
kommer fordonet att placeras inomhus på ett öppet område med färgade ljuskällor
arrangerade som hinder. Avsikten är att se skillnaderna i robotens beteende när
störningarsåsomreflektionerocholikaljuskälloriomgivningentillkommer.
Resultatet från testerna kommer att presenteras och användningen av olika sensorer
kommerattdiskuteras.
Slutsatsen är att det är en enkel och billig metod att använda ljusavkänning för en
spårningsrobot men att det framförallt bör användas som ett komplement till andra
avkänningsanordningarochintesomenfriståendemetod.
III
IV
PREFACE
First we would like to thank Baha Alhaj Hasan for the supervising and feedback
throughouttheproject.
We would also like to thank Staffan Qvarntröm for helping with the electrical and
mechanicaldesignandtothestudentassistantswhohelpedusduringtheway.
Finally, a huge thanks to Martin Åkerblad and Todd Barker for very useful and
appreciatedfeedbackonthereportandproject.
MikaelaKarlérus
BeataTörneman
Stockholm,May2016
V
VI
NOMENCLATURE
Thischapterdescribesabbreviationsusedinthisproject.
Abbreviations
LDR
LightDependentResistor
PWM
PulseWidthModulation
RGB
Red,Green,Blue
LED
LightEmittingDiode
CAD
ComputerAidedDesign
IR
Infra-Red
SDA
SerialDataLine
SCL
SerialClockLine
VIN
InputVoltage
GND
Ground
IPS
IndoorPositionSystem
GPS
GlobalPositionSystem
AIS
AutomaticIdentificationSystem
VII
VIII
CONTENTS
ABSTRACT.........................................................................................................................................................I
SAMMANFATTNING.....................................................................................................................................III
PREFACE...........................................................................................................................................................V
NOMENCLATURE.........................................................................................................................................VII
CONTENTS.......................................................................................................................................................IX
1
INTRODUCTION....................................................................................................................................1
1.1
BACKGROUND......................................................................................................................................................1
1.2
PURPOSE..............................................................................................................................................................1
1.3
SCOPE...................................................................................................................................................................2
1.4
METHOD...............................................................................................................................................................2
1.4.1 Testone..............................................................................................................................................................2
1.4.2 Testtwo.............................................................................................................................................................2
1.4.3 Testthree..........................................................................................................................................................3
2
THEORY...................................................................................................................................................5
2.1
LIGHTINTENSITYSENSING...............................................................................................................................5
2.1.1 Lightdependentresistor............................................................................................................................5
2.1.2 Photodiode.......................................................................................................................................................6
2.2
COLOURSENSING................................................................................................................................................6
2.3
MOTORCONTROL...............................................................................................................................................7
3
DEMONSTRATOR................................................................................................................................11
3.1
PROBLEMFORMULATION................................................................................................................................11
3.2
SOFTWARE.........................................................................................................................................................11
3.3
ELECTRONICS....................................................................................................................................................12
3.3.1 Microcontroller............................................................................................................................................13
3.3.2 Motordriver..................................................................................................................................................14
3.3.3 Lightsensor....................................................................................................................................................14
3.3.4 Coloursensor.................................................................................................................................................15
3.4
HARDWARE.......................................................................................................................................................16
3.4.1 Chassis..............................................................................................................................................................16
3.5
RESULTS.............................................................................................................................................................16
3.5.1 Lightsensors..................................................................................................................................................16
3.5.2 Coloursensor.................................................................................................................................................18
4
DISCUSSIONANDCONCLUSIONS...................................................................................................21
4.1
DISCUSSION.......................................................................................................................................................21
4.1.1 Lightsensor....................................................................................................................................................21
4.1.2 Colourssensor...............................................................................................................................................21
4.2
CONCLUSIONS....................................................................................................................................................22
5
RECOMMENDATIONSANDFUTUREWORK...............................................................................23
5.1
RECOMMENDATIONS.......................................................................................................................................23
5.2
FUTUREWORK..................................................................................................................................................23
REFERENCES..................................................................................................................................................25
APPENDIXA:TESTTHREE...........................................................................................................................I
APPENDIXB:THEFINISHEDROBOT.....................................................................................................III
IX
X
1 INTRODUCTION
This chapter describes in detail the thesis to be investigated, assumptions made and limitations
in the modelling and construction of the demonstrator.
1.1 Background
Self-driving cars are no longer just a surrealistic theory. In media you can follow the
developmentofself-drivingcarsthatsoonwillbeputontheroadinSwedenfortesting
purposesbyVolvo(VolvoCars,2016).CompanieslikeBMW,MercedesandTeslahave
developedself-drivingfeaturesthataresoontobereleasedonthemarketwithambition
to make fully autonomous vehicles (Business Insider, 2015). Google has since 2009
workedwiththeirself-drivingcarprojectandarerightnowtestingprototypevehicleson
theroad(Google,2016).Aself-drivingcarcanbedefinedasavehiclewithfeaturesthat
canmakeitaccelerate,brakeorsteerwithnohumaninput.Itrequiresagreatnumberof
differentsensorsasgyros,radars,GPS,tachymetersetc.andadvancedsoftwaretomake
itself-driven.Oneofthemainpurposesofaself-drivingcaristomaketheroadsaferand
facilitate daily life for commuting people. Every year approximately 1.2 million people
diesintrafficaccidents,which94%arecausedbyhumanerrors,afigurewhichcouldbe
decreasedgreatlywithuseofself-drivingtechnology(Google,2016).
Forthisthesisafullyautonomouscarwouldbetoocomplicatedtobuild,eveninasmaller
scale.Ontheotherhand,asmalleramountofsensorscouldbeusedforothertypesof
autonomousvehicles,withamissionthatissimplertopredict.Inworkingenvironments
not optimal for humans, like in mines, it could be possible to develop a much simpler
autonomousvehiclewhichcouldforexamplefollowlightinadarktunnel.Inanairport,a
simpler self-driving vehicle could be used to tow airplane when taxing on the airport
followinglightandcoloursonthegroundtosteeritswayonthefield.
1.2 Purpose
This bachelor thesis will investigate the possibility to control a robot with light only.
However,acompleteinvestigationonthesubjectwilltakelongertimeandcostmorethan
the framework and limits given so the investigation will be limited to answering the
followingmainquestion.
Whichaspectsoflightarerelevanttoatrackingrobotcontrolledbylight?
Thisquestionhasbeendividedfurtherintotwomorespecificquestions:
• Which type of light sensor is more suitable for a tracking robot and what are the
limitationsofatrackingrobotusingthistechnology?
• Isitpossibletomaketherobotobstacleavoidingbyusingcolouredlightandifso,is
thereanyspecificcolourthat’seasiertodetect?
Differentaspectsoflightcouldbeintensity,brightness,colour,wavelengthandasuitable
sensor would then be a sensor that reacts a lot to changes in one or several of these
1
aspects. Answers to above questions could hopefully be used as guide for further
developmentoflightcontrolledvehiclesorrobotstoachievedifferentbasicmanoeuvres.
1.3 Scope
TheOpen-SourceArduino-platformhascontributedinmakingtheuseofsensorseasier
andmoreapproachable.Thecodeusedforreadingthefilteredsensorvalues,red,green,
blueandcalculatingluxandcolourtemperatureiscollectedfromanopensourcelibrary.
Therefore,thetheorybehindiscoveredonlybriefly.Focushasinsteadbeenputontesting
andverifyingtheoutputfromthesensortodeterminethesensingrangeandthesampling
frequencythatmakesausefulcompromisebetweenaccuracyandbreakdistance.Also
thecommunicationbusI2CbetweenthecoloursensorandtheArduinoisfromanopen
sourcelibraryandisnotexplainedfurther.
Thisreportcoversthetheoryandconceptsbehindalightseekingandobstacleavoiding
robot.Thecalculationswillbemadebasedonindoorusagesandtherobotwillberunand
testedinanindoorenvironment.Thethesiswillonlycoverdrivingintwoenvironments,
acompletelydarkroom(0lux)andalightedroom(10lux).Theobstacleavoidancewill
be based on a colour sensor and therefore obstacles that are supposed to be avoided
needstobecolourcoded.Moreover,thereportfollowsthescopeofaBachelorthesisat
KTHandcorrespondsto10weeksofwork.
1.4 Method
To begin with a study of the equipment and the theory behind their functionality was
made,tobeabletoevaluatetherightusageofthedifferentcomponents.Theinformation
retrievalinvolvedmainlysciencearticlesanddatasheets. Forthefollowingresearch,a
complete prototype was built based on Arduino software and a robot kit. Sensors and
electricmotorcontrolwasthenaddedtotherobotvehicletogetherwithanextrapower
sourceinshapeofabatterypackage.Threedifferenttestswereconductedtoevaluatethe
sensorsandwillbedescribedbelow.
1.4.1 Testone
Acomparisonbetweentwolightsensorswasperformedbyplacingbothsensorsnextto
eachotheronabreadboardandmeasuringtheirresponsetodifferentlightenvironments.
Intheinitialstatebothsensorswerecovered,thesensorswerethenmovedfromadarker
to a brighter part of the hallway. The most suitable sensor of those two i.e. the most
sensitivesensorwaschosenforthedemonstratorandusedforthefollowingtests.
1.4.2 Testtwo
Toinvestigatethetrackingabilitiesofthechosenlightsensorthesensingdistanceand
sensingrangewereevaluated.Thetestwasdoneintwoenvironments,acompletelydark
room (0 lux) and a lit room (10 lux). The sensing distance were determined by first
measuringhowfarawaythechosenlightsensorcouldsensethelightsourceinthetwo
different environments. After that the sensors response from different distances were
conductedaccordingtofollowingsteps:
1. Measuringtheambientvalueintheroom.
2. Placingalightsource3minfrontofthelightsensor.
2
3.
4.
5.
6.
7.
Calculatingthesensorvalue.
Stepping0.5mforwardstothesensor.
Calculatingthesensorvalue.
Repeatingstep4-5untilthedistance0.5m.
Addingthevaluesinatable.
Thesensingrangetestwereconductedaccordingtothefollowingsteps:
1.
2.
3.
4.
5.
6.
7.
Measuringtheambientvalueintheroom.
Placingthelightsource1minfrontofthelightsensor.
Calculatingthesensorvalue.
Stepping0,1mtotheside.
Calculatingthesensorvalue.
Repeatingstep4-5untilthesensorvalueisbacktotheambientvalue.
Determinatethesensorsrange.
1.4.3 Testthree
A test to determine the colour sensors sensitivity to different colours were done by
placingthedemonstratoratsetdistanceseverytenthcentimetrebetween1-1.5metres
from a colour lighted obstacle. The illuminated obstacles are of own design with RGB
LED:splacedonabreadboard.Thetestwasdoneinacompletelydarkroomandinalit
roomandobstaclesinthethreedifferentcolours,red,greenandblueweretested.The
testwasconductedaccordingtothefollowingsteps:
1.
2.
3.
4.
5.
6.
7.
Placingthelightedobjectinfrontofthedemonstratoron1.5mdistance.
Calculatingthesensorvalue.
Movingtheobject0.1mforward.
Calculatingthesensorvalue.
Repeatingstep3-5untilthedistance0.1m.
Repeatingstep1-5forthethreecolours.
Plottingthevaluesinexcel.
TheresultfromthetestwillbepresentedinSection3.5.
3
4
2
THEORY
This chapter presents the theoretical framework that the performed research is based on. The
different parts will be described separately below.
2.1 Light intensity sensing
Lightsensorisadevicethatdetectstheambientlightlevelandsendsanoutputsignal
whichvarieswiththelightintensity.Lightsensorsabsorblightenergyandreactwitha
physicalalterationinaspectrumfrominfra-redtoultravioletlightandcreateelectricity
in form of electrons (Eriksson, 2003). Light sensors are commonly known as photo
sensorsandcanbedividedintotwogroups,theonesthatconvertthelightenergydirectly
intocurrent,forexamplesolarcellsandtheonesthatchangestheirelectricalproperties
depending on the intensity of the incident light and converts the measured light to a
numerical value, for example light dependent resistor, photodiode etc. The numerical
valuethencontrolsthesuppliedcurrenttothemotors.
2.1.1 Lightdependentresistor
Alightdependentresistor,alsocalledLDR,changesitsresistivevaluedependingonthe
incidentlightintensity.Theresistancewilldecreasewithincreasinglightintensityand
viceversa.Tobeabletocalculatetheresistanceofthesensors,anotherresistorneedto
beadded.Itcanbedonebyeitherapull-downorapull-upresistor.Usingapull-down
resistor,theresistorisconnectedtothegroundasillustratedin.Apull-upresistorworks
in the same way but the resistor is instead connected to its voltage source (Mims III,
2016).Thisphenomenon,toconnectaLDRinserieswitharesistorgivesacircuitcalled
voltagedivider,Figure1.
Figure1.Voltagedividerwithapull-downresistor
WhentheresistanceofthelightisdecreasingthetotalresistanceoftheLDRandthepulldown resistor will decrease, in turn the current flow through both the resistances
increases which leads to an increase of the voltage across the fixed resistor. Once the
outputofthevoltagedividerisknown,theresistanceofthesensorcanbecalculatedusing
Equation1,
5
!"#$ = !&'
()
(* + ()
(1)
where!"#$ istheoutputvoltage,!&' istheinputvoltage,() theresistorand(* thephoto
resistorsresistance(Adafruit,2015).
2.1.2 Photodiode
A photodiode convert light into voltage in direct proportion to the light intensity and
consistsofalensthatmakessurethatthelightisfocusedonacertainmaterialinthediode
thatdetectslight(MimsIII,2016).Aphotodiodehastwolevelsofoutput,eitherit’soff,in
aconditioncalledreversebias,whichmeansthatnocurrentflowthroughthediodeoron,
when the light intensity is adequate and the diode will then allow a current to flow
(ElectronicTurtorials,2016).Therefore,auseofaphotodiodeispositivewhencontrolof
the light intensity is needed, while an LDR is preferred when varying light intensities
needstobemeasured.Oppositetothelightdependentresistorwhichreactsinvisible
light,aphotodiodeismoresensitivetolightinlongerwavelengthssuchasinfra-redlight.
2.2 Colour sensing
Toenhancetheunderstandingofhowacoloursensordeviceworks,theknowledgeof
how colours are created and how humans perceive colour needs to be shared.
Electromagneticradiationinthewavelengthfromca380-780nmaredefinedasvisible
light, though the radiation need to be reflected on a surface before colour can be
distinguished.(JohJoh,KheeBoon,&Leong,2006).Thecombinationofanobject,alight
source and an observer is what creates colour nuances, where each colour represent
differentwavelengthsasillustratedinFigure2.
Colour
Wavelength
Red
~625-740nm
Orange
~590-625nm
Yellow
~565-590nm
Green
~520-565nm
Cyan
~500-520nm
Blue
~450-500nm
Indigo
~430-450nm
Violet
~380-430nm
Figure2.Colourdiagramforthedifferentwavelengths
Coloursarerepresentedbythecombinationofthethreeprimarycolours,red,blueand
green (Color Matters, 2016). These three colours constitute the additive colour model
calledRGB.Thismodelshowshowthesethreecolourscanproduceavarietyofdifferent
coloursdependingonhowtheycombineasshowninFigure3.
6
Figure3:RGBcolourmodel
Acoloursensoriscreatedtomimicthehumaneyeasmuchaspossibleandcanadjust
differingbrightnessanddetectcolourswhichgivesahighresolutionofcolourimages.An
RGBsensorisconstructedofseveralphotodiodesbehindcolourfiltersandacurrentto
voltageconversioncircuit(JohJoh,KheeBoon,&Leong,2006).Whenlightfallsonthe
photodiodeitwillbeconvertedtoaphotocurrentandthesensorproducesanoutput.The
colourcanbedeterminedbyinterpretingthesethreevoltagesandshowsinFigure4.
RGB Colour filter
Photodiode
Current-to-voltage-converter
IR
Red
Reflective or transmitted color light
IG
Green
IB
Blue
Figure4.Coloursensor
VROUT
VGOUT
VBOUT
2.3 Motor control
Motor control is about regulating and targeting the mechanisms that are essential for
movements.Amotordriverisacontrollingdeviceusedtocontrolthespeedanddirection
7
ofthemotorsi.e.tomaketherobotturncorrectly.Anessentialcomponentinafunctional
motor driver is an H-bridge which can be seen in Figure 5, this is the component that
makesitpossibleforthemotorstoturnindifferentdirections.
S1
S3
Vin
M
S2
S4
Figure5:SchematicofanH-bridge
Thetopofthebridgeisconnectedtothepowersupplyandbottomisgrounded.S1-S4are
theswitchingelements,usuallytransistors.Thedirectionofthemotorsiscontrolledby
thecurrentflowsthroughthemotorsinthedirectiondeterminedbytheswitches.IfS1
andS4areconnectedtheleftleadofthemotorwillbeconnectedtothepowersupplyand
therightleadtotheground.Thecurrentflowsthroughthemotoranditstartsspinning
andthemotorwillgoforward.IfS2andS3areconnected,thecurrentwillflowinthe
reversedirectionandthemotorwillgobackwards.TheadvantageofanH-bridgecircuit
is the possibility to drive backwards and forwards at any speed, optionally using an
independentpowersource(McManis,2006).
Inthesamewayasthedirectionofthemotorscanbecontrolledisitpossibletocontrol
the speed of the motors. The speed can be controlled with a Pulse-Width-Modulation,
PWM, which is a method to create a continuous variable power supply. Instead of a
continuouslyvaryinganaloguesignal,aPulse-Width-Modulationsignaldeliversenergy
throughasectionofpulses.Digitalcontrolisusedtocreateasignalbetweenonandoff,
somethingcalledsquarewave(Arduino,2016).ThecontrolofthePWMsignalisdone
withtwoparameters,theclockcycleandthedutycycle.Theclockcycleisthefrequency
ofthesignalandthedutycycleislinkedtotheswitchingtime.Thedurationofon-timeis
called pulse width and by increasing and decreasing the pulse width, the controller
regulatestheaverageDCvoltageappliedtothemotorthatchangesthespeedofthemotor.
Dutycycleistheproportionofon-timetotheperiodoftime.Figure6illustratesaPWM
signalanditsmeanvalue,inthiscaseisa50%dutycycle.Highmeansthatitsuppliesthe
motorwiththefullamountofavailablevoltage,whichalsocanbedescribedasa100%
dutycycle.Lowdoesn’tsupplythemotorwithvoltageatallandmeanistheaverageof
highandlow.
8
Voltage
High
Mean
Pulse Width
Period
Low
Time
Figure6:PWMsignalwitha50%dutycycle
9
10
3 DEMONSTRATOR
Thischapterpresentthedevelopingprocessofthedemonstratoraswellasthefinalresult.
3.1 Problem formulation
As stated in the purpose, the scientific question to be evaluated in this project is the
possibility to control a robot using only light. This was reduced to an investigation of
whichaspectsoflightsthatareneededtocontrolthetrackingrobot.
Thedemonstratorshouldbeabletoaccuratelysteeraccordingtotheinputfromthelight
sensors and the RGB sensor i.e. to steer towards the higher light intensity and avoid
obstaclesusingtheinputfromtheRGBsensor.Theobstacleavoidancewillbebasedona
colourcodingsysteminspiredbythenavigationsystemusedatseaandtheprinciplefor
thesystemcanbevisualizedinFigure7.Whentherobotisapproachinganobstacleit
shouldperformeitheraleftorarightturnbasedontheobstaclescolour.
Figure7.Principlefortheobstacleavoidance
3.2 Software
Thesoftwarecanbedividedintwosubsystems,oneforcomparinglightintensityandone
for detecting obstacles. The demonstrator are constructed with a left and a right light
sensor,asdescribedinFigure9.Thefirstsystemisbasedontheinputfromtheleftand
therightlightsensorandthesecondsystemisbasedontheinputfromthecoloursensor.
Thelightsensorsmeasurethelightintensitylevelsandtheseslevelsarethencompared
todetermineinwhichdirectiontherobotshoulddrivei.e.thePWMsignaltoeachpairof
motors. To keep the robot from constantly turning, an additional condition has been
setup. For the robot to turn the sensor difference between the left and the right light
sensorneedstobehigherthanathresholdvaluethathasbeenexperimentallysettothe
numericalvalue45.TheRGBvaluesthatwereusedforavoidingobstaclesarecompared
toathresholdvaluethatwassetasapercentageofthemeasuredclearlightvalue.This
way the robot can drive in different light environments and still be able to sense an
obstacle.AflowchartofthesteeringsystemisillustratedinFigure8.
11
Figure8.Flowchartofthesoftwaresystem.
3.3 Electronics
Inthefollowingsectiontheelectricalcomponentsandhowtheyareconnectedwillbe
described.InFigure9anoverallconstructionofthevehicleandtherelationbetweenthe
componentsispresented.
12
Left sensor
Right sensor
Colour sensor
Right motor
Left motor
Arduino UNO
Motor driver
Right motor
Left motor
Figure9.Schematicofthedemonstratorcontrol
3.3.1 Microcontroller
ArduinoUNOisamicrocontrollerbaseddevelopmentboardwithopensourcehardware
andsoftware.Theboardisusedtotransmitinformationfromthesensorsandcontrolthe
fourDCmotorsthroughthemotordriver.Ithas14digitalI/Opinsofwhich6canbeused
asPWMoutputsand6analogueinputs(Arduino,2016).TheArduinocallstheLDRand
when it gets a new reading from the sensors it adjusts the motor speed and direction
accordingtotheprogrammedinstructions.Thespeedadjustmentsaredonebysendinga
PWMsignaltothemotordriver,whichprocessthesignalandforwardsittothemotors.
Sincethemotorsineachsetareparallel,therightandleftsidearecontrolledseparately,
whichenablesthesteering.
13
3.3.2 Motordriver
ThemotordriverchosentopowerandcontrolthemotorsisaL298NDualH-BridgeMotor
DriverShield(ArtofCircuits,2015).EachchannelontheL298Niscapableofdelivering
anoutputcurrentupto2A,thusitcaneasilydrivetwosetsofbrushedDCmotorsas
illustratedinFigure10.
M
M
L298N Driver
Arduino UNO
Power supply
+
-
ENA
IN1
IN2
IN3
IN4
ENB
AOUT1
AOUT2
VIN
GND
BOUT1
BOUT2
PWM
AN
PWM
AN
PWM
PWM
M
M
Figure10.DC-motorcontrol
Two enable inputs are provided to enable or disable the motors independently of the
inputsignals.Moreover,themotordriverhasfourpinsusedforcontrollingdirectionof
themotors,andtwopinswhereitreadsthePWMsignalthatissenttothemotors.The
motorsinthesameseti.e.thetwomotorsonthesamesideareinparallel,whichmeans
thatthePWMoutputisthesameforthetwomotorsonthesameside.Themotordriver
isconnectedtoanexternalpowersource,inthiscasea7,4Vbatterypackwiththebenefit
topowerthemotorsdirectlywithouthavingtogothroughtheArduino.
3.3.3 Lightsensor
Thelightsensorsusedinthisthesiswherethetypeofsensorsthatconvertsthemeasured
lightintensitytoanumericalvalue.Twodifferentlightsensorsofthattypeweretested
andevaluated.ThelightsensorsthatweretestedweretheLDRandthephotodiodeand
theirfunctionalityaredescribedinSection2.1.
The photodiode was of the type TSL 252R (Elfa, 2016) and shown in Figure 11. The
photodiodewasonlyusedinthetestdescribedinSection1.4.1thusnotmountedonthe
demonstrator.
14
Figure11.Photodiode
Thelightsensorsusedforthedemonstratorandtocontroltherobotsrightorleftturns
attheforwardrunaretwoLDR:softypeB906032(Elfa,2016),showninFigure12.As
mentionedinSection2.1.1thelightsensorswereconnectedtoapull-downresistorand
theinputfromthesensorsgoesthroughtheanalogueinputsA0andA1totheArduinoas
illustratedinFigure14.Thepull-downresistorwascalculatedwithEquation1to270Ω.
Togetsomedistancebetweenthesensorstheyaremountedontopofthevehiclewithan
arrangement similar to the feelers of an insect. The mounting is described further in
AppendixB:Thefinishedrobot.
Figure12.LightDependentResistor
3.3.4 Coloursensor
ThecoloursensorusedinthisprojectwasaTCS34725,asshowninFigure13(Adafruit,
2016),whichhasbothRGBandclearlightsensingelements.ThesensorfeaturesanIR
blockingfilterthatminimizestheIRspectralcomponentoftheincominglightandallows
colourmeasurementstobemadeaccurately.Sincethehumaneyecan’tseeinfraredlight,
the IR blocking filter will make the sensing more realistic. The RGB-sensor has 7 pins
labelled,LED,INT,SDA,SCL,3V3,GNDandVIN.Fortheapplication,SDA,SCL,GNDand
VIN pins are used. The serial interface SCL is used to synchronize all data transfers
between the Arduino and the rest of the components, while SDA is the data line. The
connectionsbetweentheArduinoandsensorareillustratedinFigure14.
Figure13.RGBcoloursensor
15
LDR
5V
SDA
SCL
RGB Colour sensor
Arduino UNO
LDR
A0
A1
LED
INT
SDA
SCL
GND
VIN
GND
Figure14.DescriptionovertheconnectionbetweenthesensorsandtheArduino
3.4 Hardware
3.4.1 Chassis
ThedemonstratorisbasedonafourwheeledrobotchassissuppliedbytheMechatronic
institution at the Royal Institute of Technology. The chassis features four DC motors
whichareoperatingbetween3-6V.Thecurrentpowerconsumptionofeachmotorvaries
from100mAto300mAdependingonrotationalspeedandload(Curriculum,2015).The
chassiswasbuiltwiththreelevels.Thefirstlevelispreparedforthetwobatterypackages,
a ZIPPY Flightmax 5000mAh battery pack to supply the motors and a 2200mAh USB
power bank for the Arduino. On the second level the motor driver and sensors were
mountedandtheArduinowereplacedonthetoplayer.Thelightsensorandthecolour
sensorweremountedona3Dprintedconstruction,modelledintheCADprogramSolid
EdgeTM.
3.5 Results
Thispartcoverstheresultfromthedifferentteststhatweredoneduringtheproject.
3.5.1 Lightsensors
Thischapterrevealstheresultsyieldedfromtheperformedtests,moreorlessdivided
intoparagraphsbasedontheperformedtestsdescribedinSection1.4.
The registered sensor output during the test described in Section 1.4.1 are sorted
accordingtothemeasuredluxvaluesandplottedbelowinfigure15.Thefigureshowsthe
LDRandphotodiodesresponsestodifferentlightintensitylevels.
16
Numericalsensoroutput
1200
1000
800
600
400
200
0
0
114
175
274
373
462
789 1136 1433 21791 34196
Lux
LDR
Photodiode
Figure15.Graphoverthelightsensorresponsetodifferentlightintensitylevels
FromthetestdescribedinSection1.4.2,thesensingdistanceoftheLDRwasmeasured
to7minadarkroomandtheresultsofthelightsensorsresponseispresentedinTable
1below.Themeasuredvaluesarethenumericalsensorresponsetotheflashlightputat
differentdistancesfromthesensor.
Table1.TheLDR:sresponsetoaflashlight
Flashlightdistance(m)
Noflashlight
3,0
2,5
2,0
1,5
1,0
0,5
Litroom
345
351
356
360
372
401
446
Darkroom
0
22
25
41
84
163
342
Whendoingthetestonthelightsensorsrange,theambientvaluewas348inthelit
roomand0inthedarkroombeforetheflashlightwasturnedon.Thesensoroutput
registeredatdifferentanglesarepresentedbelowinTable2.
Table2.Thenumericalvaluesonthesensingrangetest
Deviation(°)
0
11
21
30
38
45
50
54
Litroom
394
378
375
367
365
352
350
347
17
Darkroom
188
162
143
113
57
49
23
22
BasedontheresultsdescribedinTable2thesensingrangeoftheLDRwasdetermined
andareillustratedinFigure16.
7m
2m
50°
Figure16.SensingrangeoftheLDR
3.5.2 Coloursensor
TheresultfromthetestdescribedinSection1.4.3areratherextensive,whereforethe
bulkofthedatacanbefoundinappendixA.PlottedbelowinFigure17-Figure19arethe
resultsfromthetestswithred,greenandblueobstaclesconductedinalightroom.The
y-axisshowsthenumericalsensoroutput,andthex-axisshowsthedistancefromthe
sensormeasuredinmeters.
Redobstacle
250
200
150
100
50
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5
R:
G:
B:
Figure17.Thecoloursensorsresponsetoaredobstacle.
18
Greenobstacle
400
350
300
250
200
150
100
50
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5
R:
G:
B:
Figure18.Thecoloursensorsresponsetoagreenobstacle.
Blueobstacle
400
350
300
250
200
150
100
50
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5
R:
G:
B:
Figure19.Thecoloursensorsresponsetoablueobstacle.
19
20
4 DISCUSSIONANDCONCLUSIONS
This chapter summarize and analyse the previous given results with a discussion and a
conclusion.
4.1 Discussion
4.1.1 Lightsensor
FromthetheresultinSection3.5.1andthetheorybehindlightsensorsthatwascollected
duringinformationretrievalthemostsuitablelightsensorforatrackingrobotcouldbe
determinedtobetheLDR.ThegraphinSection3.5.1showsclearlythattheLDRhasa
quickerresponsetochangesinlightintensity,whichisanecessarypropertyforatracking
robot. The major difference between the two sensors is probably because the LDR
respondbettertothevisiblelightspectrumthanthephotodiodewhichrespondbetterin
theinfraredspectra.
The idea of driving the prototype in indoor light was a lot more problematic than
expected.Somethingthatnotfirsthadbeentakentoconsiderationwasthelightsensors
sensibilitytodisturbancesfromvariouslightintheroom,whichbecameaproblemwhen
tryingtosteertherobot.Inadarkroomthelightsensorsworkedasexpected,thenthe
robot could sense the light with no disturbances from other light or reflections in the
room.ThiswassolvedbyaddingaLEDstriponthefloor,withamuchbrighterlightthan
thesurroundedlightanddisturbancesfromvariouslightwerereduced.
4.1.2 Colourssensor
The results from Section 3.5.2 shows only a slightly difference between the sensors
responseinacompletelydarkroomandalightedroom.FromFigure17-Figure19itcan
bedeterminedthatthesensorsensestheredobstacleapproximatelyatthedistance0.7m,
thegreenobstacleat0.5mandtheblueobstacleat0.3m.Probablyduetothatthered
colourhasahigherwavelength,asdescribedinSection2.2,whichcontributestoalower
spreadofthecolourintheroom.Theredcolourismoreconcentratedandwillbeseenby
thecoloursensorearlier.Thebluecolourhastheshortestwavelengthofthisthreecolour
andthereforethebluecolourisdetectedfromtheshortestdistance.Thisresultedinthe
useofredandthegreenlightedobjectsasobstacles.Figure17showsthatwhendetecting
redlight,theblueandgreenvalueswereverylow,whichisgoodwhentryingtodetecting
aspecificcolour.However,thiswillgiveaquitenarrowuseofthiskindoftrackingrobot.
Tobeabletobuildacompleteconstructionthatonlyuseslightforselfdriving,agreat
varietyoflightandcolourintensityfromtheobstaclesmustbejudgedotherwiseitwill
notwork.
Themeasurementsweredonewithanondrivingvehicle.Accordingtotheresultsfrom
Section3.5.2thecoloursensorarepossibletosensethecolouredilluminatedobstacles
withinadistancelongenoughtomaketherobotabletoturnandavoidtheobstaclein
time. In practice it didn’t work as expected. With a driving vehicle a problem with the
coloursensorandthevehiclesresponsivenesstothesensorsinputoccurred.Thecolour
sensorsmeasuredthecolourdifferencestooslowanditresultedinthecrashingintothe
obstacle,sinceitwasn’treactingfastenough.Thisproblemwassolvedbychangingthe
21
colour sensors measuring speed to measure faster and the reaction time could be
reduced.
4.2 Conclusions
Thepossibilitytocontrolarobotbyusingonlylightmaynotbethemostaccurateway.It
is not impossible to follow a track but it is surrounded by certain limitations. It is
achievableiftherobotissupposedtofollowacertainlightasaflashlightoralighttrail,
but the ability to let the robot freely drive in a normally lit room and seek after the
brightestlightsourceisunattainablewithoutaddingothersensorstoavoidobstacles.
Althoughitisnotthebestpossiblewaytoconstructatrackingrobot,couldaconclusion
stillbedrawnbasedontheinformationdeterminedduringthisthesis.Forthiskindof
trackingrobot,theLDRwasselectedasthemostsuitablelightsensor.Theabilitytoget
therobottoavoidobstaclesbyusingonlycolouredlightwerepossiblewiththeuseofa
colour sensor and the colour easiest to detect was the red colour, due to its long
wavelength.
To sum this up, using light sensing for a tracking robot are an easy and inexpensive
method,butshouldbeusedasacomplementtoothersensingdevicesnotasastandalone
method.
22
5 RECOMMENDATIONSANDFUTUREWORK
Thischaptergivesrecommendationsformoredetailedsolutionsandfuturework.
5.1 Recommendations
Asdescribedinpreviouschapterthecoloursensorismountedstraightaheadandonthe
lowerpartofthevehiclewithascreenontopofthesensor.Thismeansthatthelighted
obstacles need to be placed at a certain height above the ground which might not be
practicalinreallife.Tosolvethis,moresensorswithawiderrangecouldbeadded.
Furtherthecoloursensorwassetinananglethatmightnothavebeenthemostoptimal
angleordirection.Calculationandconsiderationofthebestdirectionandangletoputthe
lightsensorsincouldbeinvestigatedfurther.
Asthecoloursensorisalsoalightsensingdevicethesurroundinglightinaroomcould
affect the sensors signal, as the light contains different colours depending on the
environment,forexamplelightinfluorescentlampshasabluerlightthanalight-bulb.
Thecoloursensorshouldthereforebecalibratedtobeabletoworkindifferentlighted
environments.Itcanbedonebyputtingawhiteandablackpaperinfrontofthesensor
andsetthesevaluesasreferencevaluestotheinputfromthesensor.Thesensorreadings
could then be compared to these values, this way environmental differences and
disturbancesfromsurroundinglightcouldbeavoided.
5.2 Future work
The demonstrator and the control system can be improved in a lot of aspects. A
construction of a tracking robot using only light is possible, but as described earlier
certain limits are required. To construct a complete tracking robot, the demonstrator
needtobeequippedwithmoresensorsandfurtherdevelopedsoftwaretoworkcorrectly.
Onemainproblemwiththedemonstratorwhendrivingindoorwithindoorlightwasits
abilityofavoidingwallsandopendoors.Implementinganultrasonicsensororinfra-red
sensorcouldbeasolutiontotheproblem.Thedemonstratorshouldalsotobeequipped
withastartandstopsystem.Inaddition,itcouldbeinterestingtofurtherdevelopthe
abilitytoregulatethespeedofthevehicle.
ThedemonstratorcouldalsobedevelopedwithanIPSsystemforindooruseorGPSfor
outdooruse,dependingonitspurpose.Withhelpofthatsystemtherobotcouldscanthe
light intensity of a defined area and position its way back to the brightest spot of the
scannedarea.It’sanideathatcouldbeusedforaself-drivingmower,chargedwithfor
examplesolarcells.Insteadofneedingahumaninputtocarrythemowertoitscharging
station,themowercouldonitsownfindthesunandchargeitselfusingsolarcells.
Tousethiskindoftrackingrobotinenvironments,notoptionalforhumans,suchasmines
orairportswouldalsorequiremoresensors.Fortheuseinamineamappingsystemand
alocalnavigationsystemwillbeneeded.Touseitonanairportabettersafetysystemis
needed, AIS system, which is a system to see others with AIS systems, and a sensing
systemforafixedreferenceintheground(RL,2016).Areferencesystemintheground
23
willgivetherobotpositioninghelpandalightedlinewillgiveexactprecision.Theconcept
withredandgreencoloursensingcanalsobeusedinshippinglanesfornavigationof
autonomousshipsandboats.
24
REFERENCES
Adafruit,2015.Usingaphotocell.Availableat:
https://learn.adafruit.com/photocells/using-a-photocell.[Accessed:2016-03-09]
Adafruit,2016.Datasheet-TCS3472.Availableat:
https://cdn-shop.adafruit.com/datasheets/TCS34725.pdf.[Accessed:2016-04-19]
ArduinoUno,2016.Arduino–ArduinoBoardUno.Availableat:
https://www.arduino.cc/en/main/arduinoBoardUno.[Accessed:2016-04-10]
Arduino,2016.Arduino–PWM.Availableat:
https://www.arduino.cc/en/Tutorial/PWM.[Accessed:2016-03-09]
ArtofCircuits,2015.Datasheet-L298N.Availableat:
http://www2.st.com/content/ccc/resource/technical/document/datasheet/82/cc/3f/
39/0a/29/4d/f0/CD00000240.pdf/files/CD00000240.pdf/jcr:content/translations/en.
CD00000240.pdf.[Accessed:2016-03-20]
BusinessInsider,2015.10millionself-drivingcarswillbeontheroadby2020.Available
at:http://www.businessinsider.com/report-10-million-self-driving-cars-will-be-on-theroad-by-2020-2015-5-6?IR=T.[Accessed:2016-04-19]
ColorMatters,2016.ColorSystems–RGB&CMYK.Availableat:
http://www.colormatters.com/color-and-design/color-systems-rgb-and-cmyk
[Accessed:2016–04-17]
Curriculum,2015.DC-motors.Availableat:
http://curriculum.vexrobotics.com/curriculum/speed-power-torque-and-dcmotors/dc-motors.[Accessed:2016-05-08]
ElectronicTutorials,2016.Lightsensors.Availableat:
http://www.electronics-tutorials.ws/io/io_4.html.[Accessed:2016-03-21]
Elfa,2016.Datasheet-B906032.Availableat:
https://www.elfa.se/Web/Downloads/_t/ds/photocells_eng_tds.pdf?mime=application
%2Fpdf.[Accessed:2016-03-20]
Elfa,2016.Datasheet-TSL252R.Availableat:
https://www.elfa.se/Web/Downloads/_t/ds/tsl250r2r_eng_tds.pdf?mime=application%2Fpdf.[Accessed:2016-02-09]
Eriksson,Patrik,2003.Ljus-ochbildsensorer.Availableat:
http://www8.tfe.umu.se/courses/elektro/FSE/Kompendier/ljusochbildsensorer.pdf.
[Accessed:2016-03-09]
Google,2016.GoogleSelf-DrivingCarProject.Availableat:
https://www.google.com/selfdrivingcar/.[Accessed:2016-04-19]
25
JohJoh,KheeBoonandLeong,2006.Usecolorsensorsforprecisemeasurement.
OptoelectronicProductsDivisionAvagoTechnologies
MimsIII,ForrestM.AmateurScientist:ExperimentingwithLightandDarkSensors.
Availableat:http://makezine.com/projects/make-38-cameras-and-av/light-and-darksensors/.[Accessed:2016-03-20]
McManis,Chuck,2006.H-Bridge:TheoryandPractice.Availableat:
http://www.mcmanis.com/chuck/robotics/tutorial/h-bridge/.[Accessedat:2016-04-
17]
RL,2016.AIS(AutomaticIdentificationSystem)Decoding.Availableat:
http://rl.se/ais_eng.
[Accessed
at:
2016-05-07]
VolvoCars,2016.Autonomousdrivingexplained.Availableat:
http://www.volvocars.com/intl/about/our-innovation-brands/intellisafe/intellisafeautopilot/this-is-autopilot/autonomous-drive-in-detail.[Accessedat:2016-04-17]
26
APPENDIXA:TESTTHREE
TheresultfromthetestinthedarkroomdescribedinSection1.4.3areplottedbelowin
FigureA1-FigureA3.
Redobstacle
200
180
160
140
120
100
80
60
40
20
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5
R:
G:
B:
FigureA1.Thecoloursensorsresponsetoaredobstacleinadarkroom.
Greenobstacle
350
300
250
200
150
100
50
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5
R:
G:
B:
FigureA2..Thecoloursensorsresponsetoagreenobstacleinadarkroom.
I
Blueobstacle
350
300
250
200
150
100
50
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
R:
G:
0.9
1.0
1.1
1.2
1.3
1.4
1.5
B:
FigureA3..Thecoloursensorsresponsetoablueobstacleinadarkroom.
II
APPENDIXB:THEFINISHEDROBOT
FigureB1.Frontalviewofthefinishedrobot
III
FigureB2.Topviewwithboxesshowingthedifferentcomponents.
FigureB3.Frontwithboxesshowingthedifferentcomponents
IV
FigureB4.Sideviewwithboxshowingthemotordriver
ThecomponentsinFigureB2–FigureB4areasfollows:
1.ArduinoUNO
2.LightDependentResistor
3.RGBColourSensor
4.3Dprintedconstructionforthesensors
5.L298NDualH-bridgeMotorDriverShield
V
www.kth.se