-
Algoritmi de rutare bazai pe informaiile de poziie ale nodurilor
ntr-o reea ad-hoc
Proiect realizat n cadrul cursului Reele de Calculatoare i
Internet"*Petru Andrei An 2 Master IISC
-
Introducere in retele de senzori wireless / ad-hocAlgoritmi de
rutareGeneralitatiParametrii criticiAlgoritmul A: LARAlgoritmul B:
DREAM
*
-
Ce este un senzor wireless?SoC compus din:Procesor consum
redusPutere de procesare limitataMemorieCapacitate redusaModul
radioRata de transfer micaRaza de acoperire
redusaSenzoriScalariCamere de captura, microfoaneSursa energie
SensorsProcessorRadioStoragePOWER
Fig. 1 modulele componente ale unui nodFig. 2 exemple de noduri
WSN (incapsulat si SMD)*
-
Ce este o retea de senzori wireless?Este compusa din mai multi
senzori wireless (noduri)
Proprietati:Puternic limitate dpdv energetic(compromis
performanta/autonomie)Self-management, Self-organizingScalabile
(numar mare de noduri)Heterogene (noduri
organizateierarhic,dispozitive cu diferite
capabilitati)AdaptabileSecuritate sporita
Fig. 3 retea WSN cu senzori distribuiti aleator*
-
Diferente intre WSN / WAN
Retelele de senzori sunt mai denseRetelele de senzori sunt
predispuse la erori/failuresTopologia retelelor de senzori de
schimba foarte desWSN trimite mesajele broadcast, pe cand in
retelele ad-hoc comunicarea este point-to-pointNodurile din WSN pot
sau nu sa aiba identificator unic global*
-
Aplicatii ale retelelor de senzori wireless
MonitorizareScop stiintific, aplicatii in ecologie.Informatii
spatio-temporale in timp realAccesul la zone
restrictionateSupraveghere si urmarireRecunoastereControlul
perimetruluiMedii inteligenteAgriculturaProcese industriale*
-
Aplicatii ale retelelor de senzori wirelessMonitorizare
Fig. 4 o implementare WSN in scop de monitorizare a unui
obiectiv*
-
Aplicatii ale retelelor de senzori wirelessSupraveghere si
urmarire
Fig. 5 o implementare WSN in scop militar, de recunoastere*
-
Aplicatii ale retelelor de senzori wirelessFig. 6 implementari
WSN in aplicatii consumer*
-
Algoritm de rutare -> serviciu Funcii:Defineste procedurile
si infrastructura pentru transmiterea mesajelor/datelor intre
nodurile reteleiAsigura flexibilitate si adaptabilitate reteleiUn
algoritm de rutare eficient va contribui semnificativ la: autonomia
generala a sistemului, confidentialitatea datelor transmiseFig. 7
frame-ul pachetului intr-o retea TDMAFig. 8 o retea WSN*
-
Ce folosesc informatiile de pozitie ale nodurilor in retea:
GPSR Greedy Perimeter Stateless RoutingLocation-aware long-lived
route selectionDREAM Distance Routing Algorithm for MobilityLAR
Location Aided RoutingFig. 9 GPSR (rutare geografica)Fig. 10 retea
WANFig. 11 LAR (rutarea asistata de localizare)*
-
Parametrii critici pentru algoritmii de rutare:Eficienta de
rutare (correct destination high hit rate)Evitarea buclelor de
rutareAlegerea rutelor optimeViteza de rutarePerioada mica de
convergenta a reteleiRata ridicata de transfer a pachetelor in
reteaEficienta energeticaComplexitatea implementarii
*
-
Foloseste doar informatii geo pentru descoperirea rutelor si se
bazeaza pe un protocol de rutare on-deman (gen Ad-hoc on demand
distance-vector routing)Daca emitorul cunoaste o pozitie anterioara
a receptorului cat si viteza acestuia de deplasare, el va calcula o
arie unde este posibil sa se afle receptorul in momentul actual
(expected zone)Astfel, se va minimiza efectul de flood a pachetelor
menite sa identifice rutelePachetele sunt trimise doar in expected
zoneDaca un nod din exteriorul acestei zone primeste un astfel de
pachet, il ignoraDaca nodul destinatie primeste pachetul, acesta
raspunde cu pozitia sa curenta si viteza de deplasareAtunci cand un
nod intra in retea, acesta nu cunoaste informatiile de pozitie a
celorlalte noduri, motiv pentru care nodul face fall-back si va
folosi protocolul fundamental de rutare (flood the entire
network)
*Fig. 13 LAR (rutarea asistata de localizare)
-
Pentru ca LAR sa aduca un beneficiu peste traditionalul flood,
reteaua trebuie sa fie stabila! LAR va fi eficient in cadrul
retelelor dinamice (si/sau) cu noduri care dispar, insa eficienta
va fi scazuta atunci cand apar noduri noi.
Avantaje:Se evita floodarea inutila a intregii retelePerformante
ridicate chiar si in retele puternic dinamice
DezavantajeDaca reteaua nu este stabila, algoritmul este
ineficientPerformantele sunt scazute in retelele in care fluxul de
intrare a nodurilor noi este ridicat
*
-
Fiecare nod isi cunoaste pozitiaFiecare nod isi comunica adresa
si pozitia in reteaCand se trimite un pachet, acesta este inaintat
numai pe directia nodului receptorApare efectul de
distanta:Nodurile apropiate intre ele isi trimit informatii unul
celuilalt mai des decat nodurile intre care exista o distanta mai
mareCand un pachet este transmis de la nodul A catre nodul mai
departat B, informatiile despre pozitia nodului B se detaliaza pe
masura ce pachetul se propaga in reteaCand un nod isi schimba
pozitia dez, acesta trimite mai frecvent informatii vecinilor
sai*Fig. 12 GPSR (rutare geografica)
-
Frecventa de actualizare/notificare a informatiilor de pozitie a
unui nod A catre nodul B este dependenta de distanta dintre aceste
noduri. Informatiile de pozitie constau in coordonate, ci nu in cai
de rutare!
Avantaje:Permite o scalabilitate ridicata a reteleiAlgoritmul
are o eficienta de rutare buna
DezavantajeTrebuie rezolvata problema achizitiei informatiilor
de pozitie pentru fiecare nod (prin GPS = cost ridicat dpdv al
eficientei energetice)
*
-
P. Bahl and V. N. Padmanabhan. Radar: An in-building RF-based
user location and tracking system. In Proc. IEEE INFOCOMM, paginile
775784, 2000.N. Bulusu, J. Heidemann, and D. Estrin. GPS-less low
cost outdoor localization for very small devices. IEEE Personal
Communications, 7(5):28-34, 2000. Special Issue on Smart Spaces and
Environments.S. Capkun, M. Hamdi, and J.-P. Hubaux. GPS-free
positioning in mobile ad-hoc networks. In Proc. of 34th HICSS,
volume 9, pagina 9008, 2001.B. Kusy, M. Maroti, G. Balogh, P. V
Olgyesi, J. Sallai, A. Nadas, A. Ledeczi, and L. Meertens. Node
density independent localization. In Proc. of IPSN, paginile
441-448, 2006.K. Pister L. Doherty and L. EI Ghaoui. Convex
optimization methods for sensor node position estimation. In Proc.
of IEEE INFOCOMM, paginile 1655-1663, 2001.R. Nagpal, H. E. Shrobe,
and J. Bachrach. Organizing a global coordinate system from local
information on an ad hoc sensor network. In Proc. of IPSN '03,
paginile 333-348, 2003.D. Niculescu and B. Nath. SpotON: An indoor
3-d location sensing technology based on RF signal strength.
Technical Report Report #200002-02, Department of CSE, University
of Washington, Feb. 2000.D. Niculescu and B. Nath. Ad hoc
positioning system (APS). In Proc. of GLOBECOMM, volume 5, paginile
2926-2931, 2001.D. Niculescu and B. Nath. Ad hoc positioning system
(APS) using aoa. In Proc. of IEEE INFOCOMM, paginile 1734-1743,
2003.A. Savvides, C. C. Han, and M. B. Srivastava. Dynamic
fine-grained localization in ad-hoc networks of sensors. In Proc.
of MobiCom, paginile 166-179, 2001.A. Savvides, H. Park, and M. B.
Srivastava. The n-hop multilateration primitive for node
localization problems. Mobile Networks and Applications,
8(4):443-451, 2003.A. Vargas. The OMNeT++ discrete event simulation
system. In Proc. of ESM), paginile 319-324, 2001.
*
-
*Algoritmi de rutare bazai pe informaiile de poziie ale
nodurilor ntr-o reea ad-hoc
Proiect realizat n cadrul cursului Reele de Calculatoare i
Internet"
*http://www.linuxonly.nl/docs/1/74_Geographical_routing_protocols.html*The
Location Aided Routing Protocol (LAR, [Ko00]) only uses
geographical information for route discovery. It is based on an
on-demand protocol, like AODV. If the sender knows where the
destination was at some time and it knows its speed, it can
determine in which area the destination is now. This area is called
the expected zone. LAR uses this to limit the flooding of route
discovery packets. Packets are flooded within an defined area
containing both the source and the expected zone. When a node
outside this area recieves a packet, it ignores it. When the
destination receives the route discovery packet, it returns it with
its current location and speed, which can assist in future route
discoveries.
When a node enters the network, it has no information about the
geographical position of other nodes. LAR will then fall back to
the underlying protocol, which floods the route discovery packet.
For LAR to be an improvement over flooding, the network has to be
stable. LAR will perform well for moving nodes and disappearing
nodes, but not when a lot of new nodes are added to the network.
Furthermore, connections in the network has to be stable. If nodes
connect to many other nodes for a short time, location information
will not be available or accurate enough to make use of LAR.*For
LAR to be an improvement over flooding, the network has to be
stable. LAR will perform well for moving nodes and disappearing
nodes, but not when a lot of new nodes are added to the network.
Furthermore, connections in the network has to be stable. If nodes
connect to many other nodes for a short time, location information
will not be available or accurate enough to make use of LAR.*The
DREAM protocol was proposed in [Bas98]. It assumes that each node
knows its own location. Each node then communicates its address and
location through the network. When a packet is sent, it is sent to
the direction of the receiving node.
DREAM makes use of what is called the distance effect: the
greater the distance separating two nodes, the slower they appear
to be moving with respect to each other. Neighbours close by are
frequently informed of the location of a node; nodes which are
farther away only occasionally receive this information. The
further a packet travels to its destination, the more detailed the
position of the destination becomes.
Furthermore, DREAM takes into account the mobility of the nodes.
When a node travels fast, it frequently sends its location
information to its neighbours.
DREAM is similar to HSLS in that the update frequency is
dependant on the distance to a node, only it uses coordinates
instead of routing paths. This makes it very well scalable, but
introduces the problem of determining the location for each
node.*
DREAM is similar to HSLS in that the update frequency is
dependant on the distance to a node, only it uses coordinates
instead of routing paths. This makes it very well scalable, but
introduces the problem of determining the location for each
node.**