Smarter Ci*es related R&D Projects Mounir Ghogho Interna*onal University of Rabat
Smarter Ci*es related R&D Projects
Mounir Ghogho Interna*onal University of Rabat
2
Monitoring & predic*on of solar produc*on for Moroccan solar installa*ons (MoreSolar)
• Funded by IRESEN • Started May 2014 • Dura*on: 3 years
TICLab
• Objec*ves: • Develop a solar-powered WSN • Develop a hybrid (wired & wireless)
monitoring solu*on • Develop methods to predict power
produc*on and faults using ML
WEB’ log
Internet Portal
Predic/onso1ware
Intelligent Energy: MoreSolar project
3
• UIR’ssolarsystem
o 327kW(1320PVpanels;2500m2)
o 20%reducFonofelectricitybillo Minus260tonsofCO2
MoreSolar project: pilot site 1
4MoreSolar project: pilot site 2
o 2MW(7140PVpanels,50000m2)
o 30%ofcurrentelectricityneedofKenitra’sAtlanFcFreeZone
5
– Developedasolar-poweredZigBeeWSN
– Energy-awareMACanddutycycle
• Energyconsumeddaily(Wh):
Solarpanel:5.5V–450mA,3W;7.4Wh
E(T ) = 4.973+ 0.3T
T>0.081
MoreSolar project: work progress
6
– Webinterface:
• InterfacedevelopedusingHTML,CSSandJavaScript.
• Ajaxusedfordatarequest,displayandreal-Fmeupdate
• JSONusedfordataformatforthewebpage
• Webserveronthecoordinator’sArduinoMegacard
MoreSolar project: work progress
7
– SolarpowerpredicFon
Globalmodel
Clearskymodel
StochasFcmodel
PVsystemmodel
Modelingirradiance Endogenousandexhogenousdata
LocaFon,FlFngetc.
MoreSolar project: work progress
8
Aim:RetrieveradiaFoncomponentsfromskyimagecharacterisFcs,withthegoalofirradianceforecastApproach:UsemachinelearningalgorithmswithimagefeaturesandradiaFonmeasurements
FisheyeNetworkIPDomeCamera
Imagesevery10secondsfromsunrisetosunset
GlobalhorizontalradiaFon(1ssamples)
MoreSolar project: work progress
9
Physicallayersecurityforwirelessnetworks(PHY-SEC)
• FundedbyUSArmy
• 2011-2014
Cyber-Security: PHY-SEC project
• ObjecFve:usesignalprocessingtoimprove
security(tradiFonallyahigher-network-layer
issue)atthephysicallayer(beamforming,
direcFonalmodulaFon,arFficialnoise,etc.)
• Maybeusefulforsmartgridwirelesscomms..
10
DevelopmentofpredicFontechniquesofroadtrafficinurbanareas(PreForecast)
• ObjecFve:useGPSdata,geocoding,weathercondiFons(andvideoanalyFcs)topredictroadtrafficinMoroccanciFes
IBMFacultyaward(2013)
Partners:IBM,Urbasoj,Supratour…
ITS: PreTraffic project
11
Developmentofanintegratedsystemfortrafficmanagement(TrafficMan)
• ObjecFves:• Video-basedtrafficesFmaFon
• Video-basedanomalydetecFon
• Video-basedtrafficlightscontrol
• Video-basedinfracFondetecFon
• FundedbyCNRST• StartsJan2016• DuraFon:3years
Partners:MASCiR,citycouncil,otherMoroccanuniversiFes
ITS: TrafficMan project
12ITS: HowDRIVE project
• Objectives: – Develop models for the drivers’
behaviors in different conditions
– Develop models relating the driver behavior to the crash risk.
– To develop a road safety simulator and investigate cost-effective safety improvement measures.
• Partners: SUPRATOUR, MASCiR, University of Leeds, IBM, UPM6
• SubmimedtoCNRST
How do Moroccans DRIVE? a Quantitative Behavior Analysis (HowDRIVE)
13
RFID-basedluggagesorFngsoluFon(RFID-Sort)
• ObjecFve:DevelopanddeployanRFID-basedsystemtoopFmiseluggagesorFngatMohamedVairport
Traceability and Logis*cs: RFID-Sort project
• FundedbyCNRST• Starts2016• DuraFon:3years
14Smart Agriculture: SolarSYS project • SolarPoweredSmartIrrigaFonSystem(SolarSIS)
Waterpumping
DesalinaFon(ifneeded)
Decisionsupportsystem
Soilandcrop
sensing
EnvironmentalcondiFons&weatherforecast
Solarpanels(PV,CPV,etc)
actuators
IrrigaFon,applicaFonofferFlizers
Datarecords&scienFfic
recommendaFons
WatertankFerFlizers
• SubmimedtoUSAID(invitedtosubmitthefullproposal)
Partner:
15Mobile broadband: GreenNet project
Green5GcogniFvecellularnetworks(GreenNet)
• FundedbyUSArmy• Started2014• DuraFon:3years
Partners:
Fluid/Elas/cUserCentricNetworks
CoordinaFon
NetworkDensificaFon
ResourceEfficiency
MaximizeUFlizaFon
VAS
ReconfiguraFon
UserPlaneZeroCAPEXandOPEX
Scissor’seffect
SpectralAgility
ControlPlaneHeterogenity
Big data
16Network densifica*on
Largecellcoverage Smallcellcoverage
EnablingPervasiveCommunicaFonforSmartCiFes
17
Typical user’s throughput:
R =m Bn!
"#
$
%&log2 1+
SI + N
!
"#
$
%&
B: Bandwidth n: load factor (# users/cell) ( with smaller cells) m: spa*al mul*plexing factor ( with MIMO) S: received signal power ( Tx power with smaller cells) I: aggregate interference ( with smaller cells) N: noise power (masked by I with smaller cells)
Network densifica*on Poten*al
18Network densifica*on How to model base sta*on distribu*on?
TradiFonalmodeling
19Network densifica*on How to model base sta*on distribu*on?
20
Tradi*onal grid model Random spa*al model (PPP)
Network densifica*on How to model base sta*on distribu*on?
An emerging tractable approach: Stochas*c Geometry
21
Clark-Evanstest
Network densifica*on PPP: par*cularly useful for dense networks
22Network densifica*on PPP-based performance provides a lower bound for the coverage
Analogy: Rayleigh fading model provides a lower bound for coverage/link budget
23Network densifica*on Backhaul challenges
§ Backhaul networking • Many outdoor small cells sit on lampposts and street furniture,
so bulk of deployment need microwave (including mmwave) links • Absence of direct connec*on to a central aggrega*on site (CAS)
• So, because of real estate issues and connec*vity issues, small cells will mainly be connected to each other before reaching the
CAS. So, aggrega*on needed along the way, not just at CAS. § Synchroniza*on
§ Interference management
§ Backhaul bodleneck
24Cloud Radio Access Networks
25
Cloud Radio Access Networks
• OndemandcapacityaddiFon;• Energysaving:
• CentralizedcoordinaFon;• Centralizedloadbalancing;• CentralizedSleepscheduling;
• BBUpoolcentralaircondiFoning
• EfficientcoordinaFon:• Interference;• MobilityandHandover;
26IEEECommunicaFonMagazine,July2015
Naturalvs.SynthesizedSources
27IEEECommunicaFonMagazine,July2015
28
29
Network Self-Sustainability?
10-12hours
30
• CRs with energy harves*ng →Promising solu*on for connec*vity in smart ci*es.
• Opportunis*c energy harves*ng is key enabler for “green communica*on”.
• Both temporal and spa*al dynamics of the solar energy field and the mobile user traffic are cri*cal in shaping the network-wide energy requirement.
• From the case study, a metro-cellular network is self-sustainable in terms of energy for around 3-12 hours of a day depending upon the *me of the year.
• The dynamics and randomness in energy state can be exploited in future to adain energy aware load balancing and interference coordina*on.
Conclusions