Mobile Sensor Approaches to Adjust Sensor Placement

Document Sample
Mobile Sensor Approaches to Adjust Sensor Placement Powered By Docstoc
					Research On Sensor Nets



          伍民友

    上海交通大学计算机科学与工程系



          2012/7/3
     Outline
   An Introduction
   Surveillance and Coverage
   Mobile Sensors
   Reality and Terrain
   Scalability and Heterogeneous Sensor Nets
   Multiple Sensor Nets and InterSensorNets
   Semantic Sensor Nets
   Detection and Anti-detection
   Future works


                    2012/7/3                2
     Outline
   An Introduction
   Surveillance and Coverage
   Mobile Sensors
   Reality and Terrain
   Scalability and Heterogeneous Sensor Nets
   Multiple Sensor Nets and InterSensorNets
   Semantic Sensor Nets
   Detection and Anti-detection
   Future works


                    2012/7/3                3
    Trend of Computers

   Mainframes
   PCs
   Laptops
   PDAs
   What is next?



                2012/7/3   4
    Ubiquitous Computing

   Mark Weiser:
    The most profound technologies
    are those that disappear.
    They weave themselves into the
    fabric of everyday life until they
    are indistinguishable from it.


                   2012/7/3              5
        Sensory Computing

   Sensor network (sensor net)
    that monitors the earth
       千里眼,顺风耳
   Establishing a connection
    between virtual world and
    physical world



                  2012/7/3        6
What is a Sensor Net

   A collection of many
    wireless sensors
       Sensory
       Wireless communication
       Ad hoc




                  2012/7/3       7
Sensor Nets



?!
     gateway                        Sensory data



                                                   Sensory data
               Sensory data

                                                                  Sensory data


                                   Sensory data




                              2012/7/3                                      8
    Enabling Technology

   Sensors
   Wireless
   Micro-fabrication and integration
   Embedded systems




                  2012/7/3              9
What a Sensor Net Can Do
   Monitoring
       People
       Vehicles
       Animals
       Environment
       Fabs
       Supply chains
       …


                    2012/7/3   10
    Advantages of Sensornet
   Energy advantage
       Short-range communication

         Psend(Nr)


         Psend(r)
                              …      Preceive
                     r
                         Nr


       Trade-off: # of hops

                          2012/7/3              11
        Advantages of Sensor Net

   Detection advantage
       More sensors, shorter distance to
        the target

       Better signal-to-noise ratio (SNR)
                       Preceive
           SNRr  10log          10 log Psource  10 log Pnoise  20 log r
                       Pnoise
                                                r
            snr  SNR rk  SNRr  20 log            10 log k
                                                r
                                                 k


                                       2012/7/3                               12
    Advantages of Sensor Net

   Structure advantage
       Robustness due to redundancy

   In-network processing




                     2012/7/3          13
    Disadvantage

   Small sensors
   Limited on:
       Power
       Communication range & bandwidth
       Computation power
       Memory



                     2012/7/3             14
    Power

   Short-range communication
   On-demand sensory
   Short duty cycle
   More computing,
    less communication
   Field charging

                 2012/7/3       15
    Communication, Computation
    and Memory

   Scalability
   In-network processing or
    aggregation
   Efficient algorithms
       Less computing cycles
       Less storage requirement


                    2012/7/3       16
Satellite and Sensor Net
   Satellite
       Global view
   Sensor net
       Detailed view




                        2012/7/3   17
        Applications

   Military battlefield awareness
       Rapid deployment

       Monitoring battlefield or bases




                       2012/7/3           18
  Military Affairs



  Sensor systems




Deployment         Information gathering       Enemy detection

                                    2012/7/3                     19
        Applications

   Disaster situations
       Fire
       Chemical leakage
       Nuclear plant
       …



                     2012/7/3   20
Disaster and Security




                        National key project-
                    Sanxia embankment inspection




         2012/7/3                          21
        Applications

   Environmental monitoring
       Pollution
       Water level
       …




                      2012/7/3   22
Environmental Monitoring



Mine explosion accident scene              Sensor communication detecting in miner




              Huaihe River valley surroundings supervising
                                2012/7/3                                         23
        Applications

   Agriculture
       Monitor
            Temperature
            humidity
            plant disease
            insect pests




                             2012/7/3   24
Agriculture




 Water resource management


                       2012/7/3   25
Applications
   Asset and warehouse management
   Tracking assets
   With RFID




                 2012/7/3            26
        Applications

   Industrial process control
       Monitoring manufacturing processes
       Monitoring equipments, such as
        motors, pipes
       Monitoring servers
       Monitoring any irregular situations
        and failures

                      2012/7/3            27
        Applications
   Building monitoring and control
       Low wiring cost
       Cut down energy cost
       Air control
       Biological agents or chemical
        pollutants



                      2012/7/3          28
Applications
   Security and
    surveillance
       Airports
       Subways
       Nuclear power
        plants




                    2012/7/3   29
    Applications
   Health care
       Monitoring vital
        signs of patients
       In-home elder care:
        monitoring subtler
        behaviors of an
        elders




                        2012/7/3   30
        Applications

   Automotive
       Vehicle-centric sensor net
       Robot with sensors
       Non-traditional
           Mobile
           Possible long distance



                          2012/7/3   31
    Examples
   Great Duck
    Island
       190 sensors
        monitor the
        nesting petrels
       Avoid to disturb
        the bird




                       2012/7/3   32
        Examples

   BP, an oil and gas giant
       Losing track of 900 railcars in 2003
       Solution: to fit each railcar with a
        network of sensors --- that not only
        track where the railcars is at all
        times, but can monitor weight,
        internal temperature, etc.


                       2012/7/3                33
        Examples

   BP
       Monitoring customers’ liquefied
        petroleum gas fuel tanks
       25% improvement in efficiency
       The tanks have gauges but
        customer would forget to look at
        them until it was too late


                      2012/7/3             34
        Examples

   BP
       Tracking lone workers
       To avoid hazardous areas where
        they may be at risk




                     2012/7/3            35
        Examples

   Intel
       Now, all equipments are manually
        inspected once a quarter, found
        80% of problems
       With sensor net, 95% of the
        problems can be found



                      2012/7/3             36
        Examples

   In Pickberry
    vineyard, a
    sensor net
    monitors:
       soil moisture
       rainfall
       wind
                         Vineyards
        velocity and     BC
        direction


                        2012/7/3     37
    Trend of Sensor Net
    Research

   Application-driven
       Most basics are researched
       New topics from real-world
        applications




                     2012/7/3        38
     Outline
   An Introduction
   Surveillance and Coverage
   Mobile Sensors
   Reality and Terrain
   Scalability and Heterogeneous Sensor Nets
   Multiple Sensor Nets and InterSensorNets
   Semantic Sensor Nets
   Detection and Anti-detection
   Future works


                    2012/7/3               39
Surveillance
   Points of interest or
     areas of interest
   Deployment
   Structured or
    unstructured




                  2012/7/3   40
Coverage and Connection

   Communication-to-Coverage
    Ratio (CCR)
       Equivalence: 2 or 1.732?




             2                1.732




                   2012/7/3           41
Coverage and Connection




Sensing range
Communication

                2012/7/3   42
Coverage and Connection

  CCR         high            low
          Good coverage Connected
 Dense    Good commu-
             nication         Disjoint


          Connected
 Sparse                     Isolated
                 Disjoint




               2012/7/3                  43
Coverage

   0-1 coverage
       Coverage percentage
       Average walking distance
   Probability coverage
       Detection probability




                   2012/7/3        44
Coverage
   0-1 coverage




Average walking distance


                   2012/7/3   45
Coverage
   Probability coverage




                 2012/7/3   46
Exposure
   Path-based metrics
       a start point and a destination
   Best path and worst path
   A probability approach with Voronoi
    diagram




                     2012/7/3             47
Voronoi Diagram




          2012/7/3   48
Sparse Sensor Nets

   Point surveillance
   Low CCR

   How to communication
       Communicators?
       Long range communication?
       Mobile collectors?


                  2012/7/3          49
Holes and Islands
   Definition of holes and islands
   The effect of a hole in
    communication and coverage
   Communication island
   Use mobile sensors
       fill holes
       connect islands



                    2012/7/3          50
Dense Sensor Nets

   Active and sleep
   Maximum coverage
   Minimum and balanced energy
    dissipation




              2012/7/3        51
    The ACOS Algorithm
   A precise energy-aware coverage control protocol



                          s3
                                          s2

                    s4


                         s0
                                               s1
               s5




                                                    s1
                               2012/7/3                  52
     Outline
   An Introduction
   Surveillance and Coverage
   Mobile Sensors
   Reality and Terrain
   Scalability and Heterogeneous Sensor Nets
   Multiple Sensor Nets and InterSensorNets
   Semantic Sensor Nets
   Detection and Anti-detection
   Future works



                    2012/7/3               53
    Mobile Sensors
   Attached or active
   Attached mobile sensors
       like mobile ad hoc
        networks
       vehicles
       robots
   Active mobile sensors
       move or fly?




                             2012/7/3   54
    Active Mobile Sensors

   Between sensors and actuators
   For tasks related to sensor nets
       fill holes or connect islands
       collect sensory data
       charge sensor nodes
       …



                     2012/7/3           55
Mobile Sensors

   Collect sensory data
       for sparse sensor nets
       can charge sensor nodes at the
        same time
   Charge sensor nodes
       distance charging
       infrequently



                    2012/7/3             56
  Mobile Sensors

For coverage ---
  connect islands first then fill holes




                  2012/7/3                57
    Mobile Sensors

   Algorithms based on global info
       Centralized algorithm

   Algorithms based on local info
       Distributed algorithm




                     2012/7/3         58
     Outline
   An Introduction
   Surveillance and Coverage
   Mobile Sensors
   Reality and Terrain
   Scalability and Heterogeneous Sensor Nets
   Multiple Sensor Nets and InterSensorNets
   Semantic Sensor Nets
   Detection and Anti-detection
   Future works



                    2012/7/3               59
        Reality
   Most current solutions of sensor nets work
    well in simulation only
   We don’t know how the solutions work in the
    real world
       they can work poorly in practice
   Models of communication realities to improve
    simulation tools
   New network protocols that work in real
    world environments
   Creates deployable systems



                            2012/7/3          60
Terrain
   One type of the reality is terrain
   Terrain effect on communication
   Terrain effect on sensor mobility




                  2012/7/3               61
Digital Map for Mobile Sensors

   Preprocessing
   Find a path to destination




                2012/7/3         62
Digital Maps




           2012/7/3   63
Path to Destination




           2012/7/3   64
     Outline
   An Introduction
   Surveillance and Coverage
   Mobile Sensors
   Reality and Terrain
   Scalability and Heterogeneous Sensor Nets
   Multiple Sensor Nets and InterSensorNets
   Semantic Sensor Nets
   Detection and Anti-detection
   Future works



                    2012/7/3               65
Scalability of Sensor Nets
   Sensor nets are ad hoc networks
   Scalability problem
       better than general ad hoc networks
          amount of nonredundant

           Information is O(log N )
          capacity of ad hoc networks is

           O( N )
       but, it requires eliminating all
        redundant information

                      2012/7/3                66
    Scalability of Sensor Nets
   In the case of an event, amount of
    information can significantly increase
   Many hops make information transmission
       large delay
       unreliable
   Reduction of number of hops makes a
    sensor net scalable
   Solutions:
       heterogeneous sensor nets
       multiple communication channels
       directional antennas



                         2012/7/3             67
    Heterogeneous Sensor Nets
   Sensor nodes of different power and
    communication ranges
   Super nodes reduce the number of
    hops, forwarding data more efficiently
   Need flexible communication scheme




                   2012/7/3             68
Communication
   Increasing power
   Multi-channel communication
       bandwidth increase up to 3—4 times
   Multiple interfaces
   Directional antennas




                    2012/7/3                 69
     Outline
   An Introduction
   Surveillance and Coverage
   Mobile Sensors
   Reality and Terrain
   Scalability and Heterogeneous Sensor Nets
   Multiple Sensor Nets and InterSensorNets
   Semantic Sensor Nets
   Detection and Anti-detection
   Future works



                    2012/7/3               70
    Multiple Sensor Nets

   Phenomenon 1 --- battlefield:
       Air Force, Army and Marine deploy their
        own sensor networks
       Each of them may also deploy different
        networks for various sensing purposes:
            moving targets
            radiation
            light
            pressure




                              2012/7/3            71
Multiple Sensor Nets

                                           WSN
Dynamically deployment                    GATEWA
                                            Y
 System scalable
  Information sharing
                                       cooperative
   Robustness

                    WSN
                   GATEWA
                                                     Air force sensornet
                     Y




         Army sensornet



                            2012/7/3                                 72
    Multiple Sensor Nets

   Phenomenon 2 --- agriculture:
       Different sensor nets could be
        deployed:
           humidity
           temperature
           plant disease
           insect pests



                            2012/7/3     73
    Heterogeneous Net for
    Agriculture




Temperature and humidity

              Diseases and insect
   cooperation


  Basicity

                                    2012/7/3   74
    Multiple Sensor Nets
   Many sensor networks, each of different
    sensor types, may interfere each other,
    decreasing the signal-to-noise ratio (SNR)
   It is possible to change the drawback to an
    advantage
      cooperation of multiple sensor nets

   Inter-operability between sensor
    networks enables an InterSensorNet


                      2012/7/3                75
    InterSensorNet
   An InterSensorNet is composed of
    multiple sensor networks
       Overlap or partially-overlap
       relay data for each other
       process data for each other
       possible interference reduced
       energy dissipation reduced
       connectivity and robustness improved



                      2012/7/3                 76
InterSensorNet
   Topology control
   Localization and location service
   Synchronization
   In-network processing




                 2012/7/3               77
Terminology
   Local node --- node in a local
    sensor net
   Foreign node --- node in a foreign
    sensor net




                  2012/7/3               78
Topology Control
   Topology control can be more
    efficient with cooperation of
    sensor nets
   Better topology can be formed
    with foreign nodes
   Energy saving with help from
    foreign nodes



                 2012/7/3           79
Topology Control
   When a sensor node cannot find a
    neighbor in its local network, it may
    find a neighbor in another network
   Two nodes may be far apart and a
    higher communication power is
    required
       foreign nodes can help to reduce the
        power dissipation


                     2012/7/3                  80
Localization and Location
Service
   With the help of foreign nodes, a
    node is able to determine its
    location more accurately
   It can obtain a location service
    from foreign sensor nets




                  2012/7/3              81
Synchronization
   With the help of foreign nodes,
    time synchronization can be more
    accurate
   Sometime, a node cannot receive
    synchronization signal without
    help from foreign nodes




                2012/7/3               82
In-network Processing
   Some foreign nodes may be in a
    better location to process data
   Some foreign nodes may have
    necessary information for data
    processing
   Some foreign nodes may have
    higher capacity to process data



                 2012/7/3             83
Interoperability and Standard
   internetworking protocols to
    enable the cooperation among
    multiple sensor nets
   Standard of internetworking
    protocols




                2012/7/3           84
Incentives
   Why should a node deplete its
    battery doing something for
    others?
   Mechanism for incentives enables
    cooperation
   When a mechanism is available,
    nodes that serve as ``good
    citizens'' could be rewarded



                2012/7/3               85
     Outline
   An Introduction
   Surveillance and Coverage
   Mobile Sensors
   Reality and Terrain
   Scalability and Heterogeneous Sensor Nets
   Multiple Sensor Nets and InterSensorNets
   Semantic Sensor Nets
   Detection and Anti-detection
   Future works



                    2012/7/3               86
    Interoperability
   Current trend
       interconnection of multiple sensor nets
       multiple usage
   No general scalable model for
       large-scale
       heterogeneous
       multiple sensor nets
   Only a few standards for interoperability of
    sensor nets
       802.15.4 and ZigBee



                         2012/7/3                 87
    Semantics
   Semantics enables interoperability
    of multiple sensor nets

   Semantics is a glue of multiple
    sensor nets




                   2012/7/3              88
    Semantic Tags
   The current sensor nets have some
    semantic tags
       location, time, type, etc.
   The problems
       fragmented semantic tags
       implicit semantics
   No standard for semantic tags to
    support semantics of sensor nets


                       2012/7/3         89
       Semantic Sensor Nets
   A Semantic Sensor Net
    (SSN) is a heterogeneous
    sensor network that
    enables dynamic tagging
                                    Synthesis                  Synthesis
                                              New data                   New data
                                    semantics                  semantics
    of semantic information to
    sensory data and creates       Information
    semantics in the process      extraction and
    of aggregation and                                  Sensory
                                    processing Semantics data                      Sensory
    abstraction of sensory                                                 Semantics
                                                                                     data
    data to allow more           Semantics
                                          Sensory

    efficient and systematic
                                            data                 Sensory
                                                         Semantics
    monitoring and handling of
                                                                   data

    the environmental
    dynamics to provide
    demanded services.


                             2012/7/3                                                 90
    Types of Semantics
   Semantics in a sensor net
       Query Semantics
       Sensory Data Semantics
       Aggregated Data Semantics




                    2012/7/3        91
    Semantic Tags for Sensory Data
   Meta data
       necessary description about the sensory
        data itself
       example: temperature type, accuracy
   Context information
       the context information in which the
        sensory data was generated
       example: location, ID, timestamp


                        2012/7/3                  92
        Sensory Data Model
   Provide a highly scalable method
                                             Self-description
    to describe different sensory data
    and its semantics. With a uniform
    framework, sensory data can
    achieve self-description.          Semantic tag Sensory data



     Time               Pressure              Target        Quantity
                Place
                        Magnetic               Velocity          Average
       Temperature
            Sound       …                         Maximum   Minimum    …
    Sensory data semantic tag           Synthesis data semantic tag

                                   2012/7/3                                93
Semantics Example

                            Pollution       Pollution
                             survey          impact




                                         Diffusion
                            Scope
              Average                       rate
             poll. degree                               Average
                              Boundary        Time      basicity




 Pollution
  degree
                 Place          Time                               Basicity
                                                        Time


                                2012/7/3                                      94
 Flow of Semantics
                  Query
                                              Result


              Semantic layer                       Result
                                                 generation
 Pre-processing



              Synthesis data semantic layer        ……

Sensor net querying                             Synthesis data and semantics



              Sensory data semantic layer          ……

                                 2012/7/3                                      95
    Semantic Sensor Nets
                                      tag    data
                                                              tag data
                                                                                      gateway
                    tag    data        tag   data
                                                                           tag data
?
!                    tag    data       tag    data
                                                        tag data
    gateway

                tag data              tag data                       tag data


                                                                                         gateway
                   tag data                                              tag data
                                             tag data


     tag data
                           tag data
                                                          tag data            tag data

                                       2012/7/3                                            96
Query, Parsing and Execution


                       Query           SSNQL


                       Parsing


                 Executable Program    SSNEL



           SSN
SSNDL
        Description              SSN




                      2012/7/3                 97
     Outline
   An Introduction
   Surveillance and Coverage
   Mobile Sensors
   Reality and Terrain
   Scalability and Heterogeneous Sensor Nets
   Multiple Sensor Nets and InterSensorNets
   Semantic Sensor Nets
   Detection and Anti-detection
   Future works



                    2012/7/3               98
Taxonomy

   Intelligent or Non-Intelligent
    Targets (I-target or NI-target)
   Moving or stationary targets
   Target-first or sensornet-first




                2012/7/3              99
Detection
   Points or areas
   Structured or unstructured
   Measures of detection
       Coverage or exposure
       Coverage percentage
       Average walking distance
       Detection probability


                  2012/7/3         100
Detection

   Sensor net detects target
   Target detects sensor net
   Sensor net avoids target’s
    detection
   Sensor net detects target’s
    detection


               2012/7/3           101
Anti-detection
   Sensornet-first
       look around then sneak in
       find a route with smallest probability to
        be detected
       find the Voronoi diagram at real-time
   Target-first
       keep silence
       escape
       Interfere or destroy surrounding sensor
        nets

                      2012/7/3                 102
Hiding Sensor Nets
   How to reduce the probability to be
    detected by the target?
       Minimal radiation --- especially in the
        initialization phase
       Faster initialization
       Lazy report of detection results




                      2012/7/3                    103
    Other Detection Issues
   Determine I-target or NI-target
       I-target: finding a good route
       NI-target: brownian motion or straight
        line

   Intrusion detection
       When a cluster of sensors fails or
        interfered, it is possible an I-target



                        2012/7/3                 104
     Outline
   An Introduction
   Surveillance and Coverage
   Mobile Sensors
   Reality and Terrain
   Scalability and Heterogeneous Sensor Nets
   Multiple Sensor Nets and InterSensorNets
   Semantic Sensor Nets
   Detection and Anti-detection
   Future works



                    2012/7/3               105
    Future Works
   Structured sensor networks and
    deployment
   Actuators
   Dual usage of sensor nets and low-
    energy communication
   Extracting info from noisy environment
   Data mining
   Applications

                   2012/7/3            106

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:6
posted:7/4/2012
language:
pages:106