Sensor net


Published on

Published in: Technology, Business
1 Like
  • Be the first to comment

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Sensor net

  1. 1.  Large scale sensor networks are only recently emerging with alarge spectrum of applications . Distributed relaying will be shown to decrease the powerconsumption per relaying sensor node.Figure 12: Distributed relaying sensor networkfor fire detection in forests.
  2. 2. © 2003 –2004 byYu Hen Hu 3 Smart sensors Transducers Power On-board processor, storage Wireless transceivers Ad hoc network No predefined, fixed networkconfiguration Transmit, receive, and relayinformation Wireless communication Radio, infrared, optical, andother modalities Vision Smart environment:▪ Monitoring▪ Control, interaction Large number of low costsensor nodes deploy-n-play,self-configuration to formnetwork, Collaborative in-situinformation processing Applications Environmental monitoring Civil structure/earth quakemonitoring Premises security Machine instrument diagnosis Health care
  3. 3.  wireless sensor prototype byWang et al.(2005).
  4. 4. 5Wireless Sensor NetworksSensing andProcessing UnitWirelessTransceiverAd Hoc NetworkTopologyBattlefield surveillance, disaster relief, bordercontrol, environment monitoring, … etc.
  5. 5. 61,1xManagernode2,1x 1,2x 2,2x1,3x 2,3x• Data fusion– Feature vectors from different nodemeasurements are combined– Higher computational burden sincehigher dimensional (vector) data isjointly processed– Higher communication burden– Larger training data requirement1,1xManagernode2,1x 1,2x 2,2x1,3x 2,3x• Decision fusion– Decisions (hard or soft) based onnode measurements are combined– Lower computational burden sincelower dimensional data (scalardecisions) is jointly processed– Lower communication burden– Lower training data requirement
  6. 6. Two ExtremesdestinationCode and communicate all data to acentral point for processing andanalysisLocal processing and communicationbetween nodes; communicate resultto central pointdata estimatedestination
  7. 7. Physical LimitationsLow density network• low bandwidth/energyconsumption• low spatial resolutionHigh density network• high bandwidth/energyconsumption• high spatial resolution
  8. 8. KeyQuestionsHow dense should we sample ?What are the transmission rate limitationsfor a given network density ?What are the energy/power requirementsfor a given network density ?What accuracy is achievable under bandwidthand energy constraints ?
  9. 9. Signal + Noise Modelnoiseless field noisy sensormeasurements1. D(n): Achievable accuracy using n sensors ?2. E(n): Energy required to transmit data or estimates ?3. How do accuracy and energy scale with node density ?
  10. 10. MSE-EnergyAnalysishigher densityhigher resolutionmore averaginghigher densitymore datamore communication
  11. 11. Ex: Estimating a PiecewiseConstant Field
  12. 12. Estimationn sensorseach makes a noisymeasurement of thefield at its location(e.g., each contaminatedwith Gaussian noise)
  13. 13. Energy and CommunicationGoal: transmit a goodestimate of field to upperleft corner via multi-hopcommunication
  14. 14. HierarchicalComm and Data Processing• hierarchical pyramid structure for sensor networkcomm and data handling (Ganesan, Estrin, Heidiman ’02Madden et al ’02, Hellerstein et al ‘03)
  15. 15. HierarchicalCommunication
  16. 16. Hierarchical Data Fusion
  17. 17. Estimation in Action
  18. 18. FIGURESensor network protocol stack. (Reprinted from Akyildiz, I.F. et al., Computer Networks,Vol. 38, 393–422, 2002. With permission.)
  19. 19.  Sensor Nodes:sense target events, gather sensor readings,manipulate informations, send them to gateway via radiolink Base station/sink: communicate with sensor nodes anduser/operator, (database-stores the data) Operator/user: task manager, send query
  20. 20. Task Management PlaneMobility Management PlanePower Management PlaneApplication Layer: middleware, OSNetwork Layer: RoutingApplication LayerTransport LayerNetwork LayerData Link LayerPhysical Layer
  21. 21.  State of the art routing protocols aredistributed and reactive : the systems startlooking for a route only when they haveapplication data to transmit We study here Ad hoc On demand DistanceVector (AODV) and Dynamic Source Routing(DSR) for the sensor network
  22. 22.  Route DiscoveryA node sends a Route Request message to all of its neighbours.Any node receiving such a request, either answers to it or rebroadcasts it.The procedure finishes either when the request sender has received theroute information, or when the request times out. With AODV, each node remembers the next hop information associated withthe destination.The route knowledge itself is distributed in thenetwork. With DSR, the complete route is sent to the route requester. Message transmission With AODV, the message is sent to the next hop as recorded in the routingtable, and this procedure is repeated at each hop. With DSR, the message is sent with its complete route as header.
  23. 23.  Rumor Routing "Rumor Routing Algorithm for Sensor Network"by Braginsky and Estrin How to make information available in a sensornetwork Assumption: sense particular eventt whenrequested, dont know the existence or thelocation of the event
  24. 24.  An event sends out agents whichtravel the network from node tonode on a random path.Each visitleaves information about the eventin the nodes database.After apredefinedTTL the agent stops A requester also sends out an agent.After some time it will hopefullycome across the path of theinformation agent by checking thenodes databases. It can then travelthe backward references the firstagents left in the nodes to reach theevent.
  25. 25.  Critical review+ Only a small number of nodes have to adopt thesame information+ Only a small number of nodes have to process therequest When or whether requested informationcan be delivered is a random process.-The failure of nodes can interrupt the path to theevent (depending on how broad it is).-The actual behavior of a node is very different fromwhat is shown in the former slides
  26. 26.  Sensing: sensor --atransducer that converts aphysical, chemical, orbiological parameter intoan electrical signal Processing:microprocessor(CPU)data storage(Mem)AD converter Communicating: datatransceiver(Radio), Energy source: battery
  27. 27. 68HC11NodeSpecificOS ModulesMiddlewareMiddlewareSensorDriverNodeSpecificOS ModulesHardware SensorMiddlewareARMTemperatureSensorSensorDriverNodeSpecific OSModules68HC11PressureSensorNodeSpecificOS ModulesSensorDriverMiddlewareSensor NodeModel• 68HC11 µC• No Sensor• ARM Microcontroller• Temperature Sensor• 68HC11 Microcontroller• Pressure Sensor
  28. 28. Picoradio(UCB)WINS(UCLA)SmartDust(UCB)Sensor,ActuatorBatteryProcessor HF
  29. 29. Characteristics of Sensor Nodes Limited capacity of Battery (Lifetime: day - 10 years) Processing capabilities (10MHz) Transmission range (5 - 20 meters) Data rates: Bit/s - KB/s Transmission methods: 802.11 (WiFi) Bluetooth – short distance, other applications ZigBee – for sensor network Price: some cents
  30. 30.  Storage persistent storage for datastreamsIntegrity Service/Access ControlQuery ManagerStorageSensor Manager
  31. 31.  Query Manager manages active queries query processing delivery of events andquery results toregistered, local orremote consumersIntegrity Service/Access ControlQuery ManagerStorageSensor Manager
  32. 32. Integrity Service/Access ControlQuery ManagerStorageSensor Manager Top layer: accesscontrol and integrityservice OS examples: TinyOS: when an eventoccurs, it calls theappropriate eventhandler to handle theevent. Others: Contiki,MANTIS, and SOS.
  33. 33.  Create Hardware-optimized softwarecomponents (driver, operating system ) Create hardware- independent softwarecomponents (middleware, services) Combining of predefined components Source code generation Removing unused components Optimizaion of interface Optimizaion to nodes hardware Distribution of nodes in different environments Monitoring the execution Creation of logfiles Evaluation of logfilesComponentsDesign & EditComplie/LinkDistributeExecute/AdministrateEvaluationResourceHardwaredrivenMonitoringOptimization
  34. 34. www.themegallery.com2:Securityapplication4:MedicalApplication1: Militaryapplications3:Environmental application5:Commercial applicationLocation of combatants,vehicles and weaponson the battlefieldIt may be used to monitorpatients from a distance•Report a possibleoutbreak of fire•Detect dry areasDetect movements of the earthto predict earthquakeFacilitate stockmanagement16/24
  35. 35. 17/24
  36. 36. 38
  37. 37.  Based on the IEEE 802.15.4 Standard Popular for WSN devices. ZigBee adds: Network topologies Interoperability with other wireless products39
  38. 38.  TinyOS is a free and open source operating system. TinyOS is an embedded operating system written in thenesC.40
  39. 39.  Given: Field A N sensorsHow well can the field be observed ? Closest Sensor (minimum distance) only WorstCase Coverage: Maximal Breach Path Best Case Coverage: Maximal Support Path Multiple Sensors: speed and path consideredMinimal Exposure Path
  40. 40.  Applications of Wireless Sensor Networking:In the present era there are lot of technologies which areused for monitoring are completely based on the wirelesssensor networking. Some of important applications areenvironmental monitoring, traffic control application,weather checking, regularity checking of temperature etc.Wireless sensor networks can also be used for detectingthe presence of vehicles such as motor cycles up to trains.These are some important wireless sensor networkingbased technologies which help us in our daily life. Some ofthere daily life applications are: used in agriculture, waterlevel monitoring, green house monitoring, landfillmonitoring etc.
  41. 41.  In studying the performance or a wireless sensornetwork, you must take into consideration thedeployment scenario which includes; topology,radio ranges, trajectory of targets and eventtraffic, and trajectories of user nodes and querytraffic. All of these affect design trade-offs, andtherefore any algorithm or protocol chosenshould be evaluated under diverse deploymentscenarios.
  42. 42. © 2003 –2004 byYu Hen Hu 45 Sensor network is a new application area forcomputer vision, graphics and imageprocessing It requires multi-modality, multimediaprocessing under the constraint ofminimizing communication and energyconsumption.
  43. 43.  Sensor Network can be used in manyapplications, such as Military, Environmentaland Health…etc. Its characteristics are tiny node, low power,limited resources, dynamic network topologyand various scales of network deployment. Middleware is used to connect the networkhardware, operating systems, network stacks,and applications in different approaches. For examples,Virtual Machine, Mobile Agent,Database and Message Oriented.
  44. 44.  Security in Sensor Networks. Public/Private Key▪ Key establishment beyond sensor network capabilities. Shared Key▪ Simple solution, but single node may reveal the secret key.▪ Scalability? → each node stores n-1 keys (n(n-1) keys need to beestablished) Solution? Privacy Aspects in Sensor Networks. Sensor technology may be used for illegal surveillance. Providing awareness of the presence of sensor nodes? Solution?
  45. 45.  FIGURESensor signal processing flow.
  46. 46. 49 ISI team experimented with three iPAQ-basedvideo sender nodes and collected videobaseline of several vehicles. RTP packet dumps andVHS video tape. VT team supported BBN integratedexperiment with Sensoria 2.0 nodes. UCLA ran developmentalexperiments on sensor fieldcoverage algorithms (underSensorware project).
  47. 47. 50
  48. 48. 51
  49. 49. 52