Linköping Studies in Science and Technology. Dissertation No PDF

Description
Linöping Stuies in Science an Technology Dissertation No. 49! # $ % & ' ( ) * + $ & ', * - ( * %. / Department of Computer an Information Science Linöping University SE Linöping, Sween Linöping

Please download to get full document.

View again

of 62
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information
Category:

Finance

Publish on:

Views: 37 | Pages: 62

Extension: PDF | Download: 0

Share
Transcript
Linöping Stuies in Science an Technology Dissertation No. 49! # $ % & ' ( ) * + $ & ', * - ( * %. / Department of Computer an Information Science Linöping University SE Linöping, Sween Linöping Copyright Eri Kuiper Cover esigne by Liana Pop UAV images on cover copyright Saab AB ISBN ISSN Electronic version available at: Printe by LiU-Tryc ii Communication is a ey enabler for cooperation. Thus to support efficient communication humanity has continuously strive to improve the communication infrastructure. This infrastructure has evolve from herals an rien couriers to a igital telecommunication infrastructures base on electrical wires, optical fibers an raio lins. While the telecommunication infrastructure efficiently transports information all over the worl, there are situations when it is not available or operational. In many military operations, an isaster areas, one cannot rely on the telecommunication infrastructure to support communication since it is either broen, or oes not exist. To provie communication capability in its absence, a hoc networing technology can be use to provie a ynamic peer-base communication mechanism. In this thesis we stuy geographic routing in intermittently connecte mobile a hoc networs (IC-MANETs). For routing in IC-MANETs we have evelope a beacon-less elay-tolerant geographic routing protocol name LAROD (location aware routing for elay-tolerant networs) an the elay-tolerant location service LoDiS (location issemination service). To be able to evaluate these protocols in a realistic environment we have use a military reconnaissance mission where unmanne aerial vehicles employ istribute coorination of their monitoring using pheromones. To be able to preict routing performance more efficiently than by the use of simulation, we have evelope a mathematical framewor that efficiently can preict the routing performance of LAROD-LoDiS. This framewor, the forwar-wait framewor, provies a relationship between elivery probability, istance, an elivery time. Provie with scenario specific ata the forwar-wait framewor can preict the expecte scenario pacet elivery ratio. LAROD-LoDiS has been evaluate in the networ simulator ns- against Spray an Wait, a leaing elay-tolerant routing protocol, an shown to have a competitive ege, both in terms of elivery ratio an overhea. Our evaluations also confirm that the routing performance is heavily influence by the mobility pattern. This fact stresses the nee for representative mobility moels when routing protocols are evaluate. This wor has been supporte by LinLab, a research center for future aviation systems, establishe by Saab an Linöping University, an the KK founation through the inustrial grauate school SAVE-IT. iii iv : ; = = A 6 9BC B5 ; 7 D B7 5 B99 A E A 5 D 9B5 C = 5 Iag förväntar vi oss att vi an omma ut på Internet nästan var som helst och omeelbart få ontat me vänner och beanta. Sälet att enna möjlighet finns är att vi har byggt upp en omfattane infrastrutur som hjälper oss att sica atapaet fram och tillbaa. Låt oss anta att enna infrastrutur inte längre finns och att et ena som finns tillgängligt är e eletronisa apparater vi bär på oss me trålösa anslutningsmöjligheter som Wi-Fi. Genom att samarbeta och sapa ett a hoc-nätver me hjälp av essa apparater är et möjligt, men långt från trivialt, att sica information mellan personer som inte an ommunicera iret me varanra. I ett a hoc-nätver ansvarar ett routing-protooll för att hitta en väg i nätveret för atapaet så e ommer fram till rätt mottagare. Routingprotoollet sicar ett paet trålöst från apparat till apparat tills et ommer fram till mottagaren. Det har oc visat sig svårt att esigna routing-protooll som fungerar bra å e personer som bär på apparaterna rör på sig. Ett annat problem är att et iblan inte finns någon väg för ett paet att nå mottagaren. Sälet är oftast att et sanas personer på strategisa platser som gör att nätveret elas upp i öar är snabb ommuniation bara fungerar inom essa öar. Man säger å att nätveret är partitionerat. Om vi äremot an täna oss att vara lite tålmoiga så an vi tillåta att tillfälligt lagra vårt paet hos en person som ommer att röra på sig och omma till ön är mottagaren finns. När etta ser så an paetet levereras, oc me en större förröjning. Den förröjning vi pratar om här är lite för lång för att fungera me ett chatt-program, men för e-post fungerar et bra. Ett sätt att göra jobbet enlare för routingprotoollet är om vi an sica paetet till en geografisa position är mottagaren befinner sig. Detta räver att alla eletronisa apparater vet var e befinner sig, men me GPSmottagare i nästan varje apparat är etta oftast inget hiner. För att veta var mottagaren befinner sig behöver routingprotoollet ta hjälp av en loalisationstjänst som an förmela enna information. v Vi har tagit fram ett routingptotooll, LAROD, för geografis routing i system me mobila noer är mobiliteten måste utnyttjas för att leverera paeten å systemet är partitionerat. För att unna leverera information om var en mottagare befinner sig har vi tagit fram en loaliseringstjänst, LoDiS. Då vi inte tror att et främst är u och jag som har behov av enna typ av routing (vi har ju tillgång till en infrastrutur), så har vi utvärerat essa protooll i ett militärt spaningsscenario. I spaningsscenariot har vi antagit att ett antal obemannae flygfaroster (UAVer) samarbetar för att spana av ett visst områe. När en UAV upptäcer något av intresse rapporterar en enna information till en annan UAV som an agera på informationen. I våra utväreringar av protoollen har vi sett att e fungerar väl och att informationen ommer it en sall (även om et tar lite ti). vi First I want to than Saab Aeronautics, an then especially Aners Pettersson an Gunnar Holmberg, for giving me this opportunity to pursue a PhD. Without the connection to a practically applicable problem omain I woul probably not have chosen octoral stuies. Tomas Jansson, my manager when this journey starte, was also important since he helpe me fin this challenge when I neee it. I also want to than my inustrial avisor Mats Eman. I might not have sought your avice that extensively, but the iscussions we ha gave me some things to thin about. I also owe gratitue to Magnus Svensson an Aners Boin, my project leaer an manager respectively, for accepting that I coul not spen all my time on the project I was part of at Saab. To my acaemic avisor an supervisor Simin Najm-Tehrani I woul lie to exten a than for guiing me to this point. You especially taught me how to write for an acaemic auience an not only to report my finings. I might not entirely agree with the anatomy of acaemic articles, but hopefully I now unerstan it reasonably well. To Di Yuan, my assisting acaemic avisor, I give my thans for helping me with the formal efinition of the forwar-wait framewor. I also want to than the other members of RTSlab for your frienship an valuable comments. You helpe me to become aware of some of the assumptions I hae mae an you pushe me to clarify an further investigate some issues. I am grateful to SAVE-IT an the KK founation for partially funing my research. You might not be in my thoughts every ay, but without you this research might never have been one. Finally I want to than my frien C. Without you I might never have selecte to wor for Saab, an then this woul never have happene. Eri Kuiper Linöping, January vii viii F Introuction.... IC-MANETs Noe Mobility Problem Description Contributions Thesis Outline... 9 Relate Wor.... Routing..... Delay-tolerant Routing Geographic Delay-tolerant Routing Opportunistic Routing Beacons-less Routing Location Services....3 Routing Moeling The ICT Moel Other Routing Moels Mobility Moels Real-Worl Mobility Moels Synthetic Mobility Moels Noe Density an Connectivity Reconnaissance Mobility Scenario The Three Way Ranom Mobility Moel The Pheromone Reconnaissance Mobility Moel Evaluation Scan Coverage Scan Characteristic Communication potential Routing in IC-MANETs LAROD LoDiS The LoDiS Protocol The Requirements on Cloc Synchronization Evaluation Scenario Parameters an Set Up LAROD Parameters LoDiS Parameters an Performance ix 4.3.4 The Impact of the Mobility Moel LAROD-LoDiS Compare to Spray an Wait Concluing Remars A Framewor for Performance Analysis The Forwar-Wait Framewor Characterizing Forwaring an Waiting Abstract Mobility an Routing Moel Framewor Inputs Base on Moels Distributions from ns- Data Characterizing Distance to Destination Valiation of the Framewor Delivery Ratio Preictions Distance an Time Preictions The Valiity of the Framewor Assumptions Practical Application of the Framewor Delivery Ratio Preictions Source Noe Parameter Ajustment Concluing remars Where We are an Where We Can Go Summary an Conclusions Future Wor Acronyms References... Appenix Percolation Theory Metrics Appenix Implementing Spray an Wait in ns Appenix 3 Complexity of Deriving the Distribution of Forwaring Distance... 4 Appenix 4 Ranom Variables Appenix 5 Distance Distributions Between Pairs of Noes x G 3 H Figure. Networ type relate to noe ensity... Figure. OR forwaring example...7 Figure 3. Forwaring areas....9 Figure 4. A taxonomy of location services... Figure 5. Change of average irection near the eges Figure 6. UAV line scanning Figure 7. Three way ranom state iagram...43 Figure 8. Local pheromone map after 36 simulate secons Figure 9. Global pheromone view after 36 simulate secons...45 Figure. Local pheromone map after 7 simulate secons Figure. Global pheromone view after 7 simulate secons...46 Figure. Pheromone search pattern...47 Figure 3. Coverage for pheromone mobility with ifferent transfer probabilities....5 Figure 4. Coverage comparison for pheromone mobility with global map....5 Figure 5. Coverage for pheromone mobility with ifferent coverage factors...5 Figure 6. Coverage for ifferent mobility moels Figure 7. Pheromone mobility. Coverage variability Figure 8. Next scan pf for ifferent mobility moels...55 Figure 9. Next scan pf for pheromone mobility with ifferent noe egrees Figure. Average number of clusters. Pheromone mobility with ifferent noe egrees...57 Figure. Pheromone mobility. Cluster variability Figure. Comparison of ifferent mobility moels...58 Figure 3. Forwaring area examples...6 Figure 4. Delay curve examples...6 Figure 5. Delay time parameters illustration...63 Figure 6. LAROD pseuo coe...65 Figure 7. LAROD-LoDiS path visualization example...66 Figure 8. LoDiS pseuo coe Figure 9. Delivery ratio with parameterize LoDiS an Pheromone mobility Figure 3. Overhea with parameterize LoDiS an Pheromone mobility.74 xi Figure 3. Average irection error at source noe Figure 3. LADOD-LoDiS elivery ratio for ifferent scenarios Figure 33. LADOD-LoDiS overhea for ifferent scenarios Figure 34. Spray an Wait elivery ratio for ifferent scenarios...77 Figure 35. Spray an Wait overhea for ifferent scenarios...77 Figure 36. Delivery ratio for ifferent pacet life times uner pheromone mobility Figure 37. Overhea for ifferent pacet life times uner pheromone mobility Figure 38. Delivery ratio for ifferent transmission loas uner pheromone mobility....8 Figure 39. Overhea for ifferent transmission loas uner pheromone mobility....8 Figure 4. Delivery ratio for ifferent noe ensities uner pheromone mobility....8 Figure 4. Overhea for ifferent noe ensities uner pheromone mobility....8 Figure 4. Time-istance forwar illustration...84 Figure 43. An illustration of minimum progress forwaring area...9 Figure 44. Probability of being forware by at least a given istance from the source with no initial wait (ccf), for four noe ensity values Figure 45. Probability of being forware by at least a given istance from the source after a wait (ccf), for four noe ensity values Figure 46. Probability of waiting for at most a given time, for four noe ensity values...98 Figure 47. Probability of being forware by at least a given istance from the source with no initial wait (ccf).... Figure 48. Probability of being forware by at least a given istance after an initial wait (ccf).... Figure 49. Probability of waiting for at most a given time before forwaring (cf).... Figure 5. Cf of source-estination istance....3 Figure 5. Preicte an simulate elivery ratio for ifferent TTLs...5 Figure 5. Probability of reaching at least 4 raio raii with respect to time....6 Figure 53. Probability of requiring at least secons with respect to istance....7 xii Figure 54. Delivery ratio for ifferent sizes of the any-to-any scenario with a constant average noe egree...9 Figure 55. Delivery ratio for ifferent sizes of the any-to-c&c scenario with a constant average noe egree... Figure 56. Delivery ratio for ifferent noe ensities of the any-to-any scenario with a constant area size... Figure 57. Delivery ratio for ifferent noe ensities of the any-to-c&c scenario with a constant area size... Figure 58. TLL require for achieving 95% elivery guarantee for various ensity levels.... Figure 59. Delivery probability in istance an TTL.... Figure 6. Pacet exchange...36 Figure 6. An illustration of the custoian location after first hop...4 xiii xiv G 3 I Table. Ranom waypoint parameters...3 Table. Gauss-Marov parameters Table 3. Connectivity properties...39 Table 4. Average noe egree in ifferent routing stuies....4 Table 5. Scenario parameters...4 Table 6. Pheromone map parameters Table 7. Pheromone parameter efinition Table 8. UAV pheromone action table...48 Table 9. Never scanne area...56 Table. LAROD parameters...63 Table. Basic simulation parameters...7 Table. Delivery ratio for ifferent LAROD parameters....7 Table 3. Overhea for ifferent LAROD parameters (transmissions per ata pacet)....7 Table 4. Key notation an terms...85 Table 5. Terms in custoian selection....9 Table 6. Probability of having an initially empty forwaring area...93 Table 7. Average probability of forwaring after a retry interval...96 Table 8. Notation an terms in eriving istributions for the waiting phase Table 9. Average wait times an forwar istances....8 Table. Percolation theory an a hoc networ parameters, an their relationships Table. Spray an Wait implementation choices an rationale Table. Impact of lost pacets on Spray an Wait...38 Table 3. Scenario parameters...39 Table 4. Spray an Wait parameters...39 xv xvi Introuction The sharing of information is vital for many tass, an the faster information can be isseminate the sooner or better a tas can be complete. With the prevalence of wireless technologies lie GSM, 3G an Wi-Fi, information is often available anytime an anywhere. The limitation of these technologies is that they require an infrastructure of base stations or access points to function. In environments such as isaster areas, or in certain war zones, this type of infrastructure is generally not available, but information exchange is still esire. In these environments you commonly use long range raios that enable point-to-point communication. The problems with these systems are that they are often expensive, buly an only provie low banwith communication. At the other en of the spectrum there are cheap, small, low power, high banwith, but short range raio technologies. If a lot of actors are equippe with this type of raios then they coul automatically form a networ an cooperate to forwar messages for each other. These types of networs that are cooperatively forme an o not rely on any infrastructure are often calle a hoc networs. To create local a hoc networs there exist technologies lie Bluetooth [93] an ZigBee [], but the creation of larger an highly ynamic a hoc networs is still in the research omain. The stuy of routing in ynamic an infrastructure-free networs starte in the fiel of mobile a hoc networs (MANETs). A ey assumption mae in MANET routing is that any two noes can always fin a connecte path to reach each other. Many MANET stuies foun that this assumption was not always vali an that a significant number of pacets were lost in the networ ue to networ partitions. About the same time other researchers were stuying how information coul be transferre in interplanetary networs (IPN) [39]. In these networs some of the communication lins are only available uring certain perios an in many cases a contemporaneous en-to-en path is never available. To cope with this they evelope the concept of custoians where the responsibility of a ata pacet is transferre CHAPTER between custoians. The problems aresse by the IPN research were also foun in other challenge networs an the fiel of elay-tolerant networs (DTNs) was efine. The main characteristic of a DTN is that it generally taes a long time for information to reach its intene estination. The reasons for the elays may be long transfer latencies or infrequent communication contacts. This means such isparate networs as IPNs, willife tracing systems [5][9] an networing in rural communities [][76] are all DTNs. In DTN routing research it is common to stuy systems where the noes most of the time have no contact with other noes, an where they only occasionally meet [94]. We call this type of DTN a solitary DTN. In this thesis we stuy networs that lie in the region between solitary DTNs an MANETS. We call this type of networ intermittently connecte MANET (IC-MANET). The characteristic of this networ type is that noes form connecte clusters, within which MANET style routing is possible. To reach noes that are locate outsie a cluster, DTN style routing is require. Due to the noe mobility, the noes that are part of a cluster constantly change. Figure provies an illustration of how the noe ensity affects the networ type. This is a very simplifie picture an it assumes that connectivity is improve with increase noe ensity. Due to noe aggregation at popular locations a networ might never be fully connecte, even if the average noe ensity is very high. Solitary DTN IC-MANET MANET Density Figure. Networ type relate to noe ensity. The communication performance in an IC-MANET is greatly affecte by how the noes move, which in turn etermines the communication opportunities that exist in the system. For this reason it is important to use a relevant scenario when a routing protocol is evaluate. The main evaluation scenario in this thesis is a military reconnaissance mission where a group of unmanne aerial vehicles (UAVs) cooperate to fulfill the reconnaissance requirement. We have also use synthetic ranom mobility moels as comparative references. INTRODUCTION 3 J K J L M N - O P Q % MANET research stuies how to route ata pacets in a networ of mobile noes with wireless raios. A ey assumption mae by all MANET routing protocols is that there exists a contemporaneous en-to-en multi-hop path between any sener an receiver, an when such a path oes not exist, they will fail to eliver messages. This oes not mean that it is impossible to route messages in the absence of contemporaneous paths, only that other principles nee to be use. In an IC-MANET the system is so sparse, or the noes are moving in such a way, that there exists at least two clusters of noes for which there is no contemporaneous path between the noes in ifferent clusters. Due to noe mobility these clusters are not stable, but instea they constantly split an merge. To enable communication between noes in ifferent clusters messages alternate between multi-hop forwaring within a cluster, an storage until the clusters reform, an further forwaring becomes possible. A consequence of the relatively slower forwaring by noe mobility compare to wireless forwars is that elivery times will be longer than if a contemporaneous en-to-en path existe. In most networs the responsibility for reliable transfer is place on the source-estination pair, an networ failures result in time-outs an resening of ata from the source. In a DTN or an IC-MANET this en-toen reliability mechanism is not a suitable esign choice since it relies on status exchanges betwee
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks