transcript
Microsoft PowerPoint - WSN&Car []Jang Ping Sheug g Dept. of
Computer Science
National Tsing Hua UniversityNational Tsing Hua University
OutlineOutline
Introduction to Wireless Sensor Networks (WSNs) Research Issues in
WSNsResearch Issues in WSNs Applications of WSNs Currently Research
Results
2007/11/8 Jang-Ping Sheu, NTHU 2
Introd ctionIntroduction
A new generation of massive-scale sensor networks suitable for a
range of commercial and e wo s su ab e o a a ge o co e c a a d
military applications is brought forth by
Advances in MEMS (micro-electromechanicalAdvances in MEMS
(micro-electromechanical system technology) Embedded
microprocessorsEmbedded microprocessors
2007/11/8 Jang-Ping Sheu, NTHU 3
Introduction
Tiny, cheap sensors may deploy onto roads, walls, or machines,
creating a digital skin that senses a variety of physical phenomena
of interest
Monitor pedestrian or vehicular traffic and intelligent p g
transportation grids in human-aware environments Report wildlife
habitat conditions for environmental conservation Detect forest
fires to aid rapid emergency responses Track job flows and supply
chains in smart factories
2007/11/8 Jang-Ping Sheu, NTHU 4
Challengesg
Limited hardware: Each node has limited processing, storage, and
communication capabilities, and limited energy supply and bandwidth
Limited support for networking: The network isLimited support for
networking: The network is peer-to-peer, with a mesh topology and
dynamic, mobile, and unreliable connectivityob e, a d u e ab e co
ect v ty Limited support for software development: The tasks are
typically real-time and massivelytasks are typically real-time and
massively distributed, involve dynamic collaboration among nodes
and must handle multiple competing events
2007/11/8 Jang-Ping Sheu, NTHU 5
nodes, and must handle multiple competing events
Sensor Node: MoteSensor Node: Mote
Mote modules were developed by U.C. Berkeley and Crossbow
Technology Inc. Power, size, and cost constrained
Slow clock cycles of microcontrollerSlow clock cycles of
microcontroller Small memory Small number of HW controllers
TinyOS
Sensor DevicesSensor Devices
Sensor component
supply
Sensor Boards for MotesSensor Boards for Motes
MTS300 Multi Sensor Board Light, Temperature Microphone, Sounder p
, Tone Detection Circuit Compatible with MICA, MICA2
MTS310 Multi Sensor Board Light, Temperature
i h S dMicrophone, Sounder Tone Detection Circuit 2-Axis
Accelerometer 2-Axis Magnetometer Compatible with MICA, MICA2
2007/11/8 Jang-Ping Sheu, NTHU 8
Research Issues in WSNsResearch Issues in WSNs
Localization and Tracking Time SynchronizationTime Synchronization
Routing Protocols Topology Control Coverage Problems Databases,
Platforms, and Tools
2007/11/8 Jang-Ping Sheu, NTHU 9
LocalizationLocalization
For several sensor network applications, including target tracking
and habitat monitoring knowing the exact location where information
was collected is critical The value of the information collected
can be enhanced if the location of the sensors where e a ced t e
ocat o o t e se so s w e e readings were made is also
available
2007/11/8 Jang-Ping Sheu, NTHU 10
LocalizationLocalization
How To Get Position? Just assume that don’t have any reason orJust
assume that, don t have any reason or, The placement of all nodes
are regular topology oror, All nodes equipped with GPS or,
Inferring positional information from some constraints or
condition
2007/11/8 Jang-Ping Sheu, NTHU 11
LocalizationLocalization
GPS build nodes localize themselves within a few meter’s accuracy
by listening to signals emitted by a number of satellites GPS’s
cons
Expensive Difficult to incorporate into every sensor nodeDifficult
to incorporate into every sensor node Power consumption
2007/11/8 Jang-Ping Sheu, NTHU 12
Range-based Schemeg
Determine the distance between two different sensor nodes based
onsensor nodes based on
Time of Arrival (TOA) or, Time Difference of Arrival (TDOA) or,
Angle of Arrival (AOA), org ( ) Received Signal Strength
(RSS).
2007/11/8 Jang-Ping Sheu, NTHU 13
Range-Based Localization AlgorithmsRange Based Localization
Algorithms
2007/11/8 Jang-Ping Sheu, NTHU 14
Range-Free SchemeRange-Free Scheme
Two kinds of sensor node Beacon nodeEquip GPS and can get ownBeacon
nodeEquip GPS and can get own location information by GPS
l d ’ i G SNormal nodeDon’t equip GPS
2007/11/8 Jang-Ping Sheu, NTHU 15
An Example of WSNs with Beacon NodesNodes
2007/11/8 Jang-Ping Sheu, NTHU 16
Example of Connection-Imposed P i it C t i tProximity
Constraints
2007/11/8 Jang-Ping Sheu, NTHU 17
Estimative RectangleEstimative Rectangle
The Estimative Rectangle (ER) of circles as a minimum rectangle
that cover the intersection of u ec a g e a cove e e sec o o these
circles. Note worthily the four edges of ER must parallel x-Note
worthily, the four edges of ER must parallel x- axis or
y-axis.
2007/11/8 Jang-Ping Sheu, NTHU 18
Example of ERExample of ER
2007/11/8 Jang-Ping Sheu, NTHU 19
The Method of Location EstimationThe Method of Location
Estimation
Regard the center of the ER as the estimative location
2007/11/8 Jang-Ping Sheu, NTHU 20
Time SynchronizationTime Synchronization
Since the nodes in a sensor network operate independently, their
clock may not be, or stay depe de y, e c oc ay o be, o s ay
synchronization with one another. This can cause difficulties when
trying to integrateThis can cause difficulties when trying to
integrate and interpret information sensed at different nodes
2007/11/8 Jang-Ping Sheu, NTHU 21
Clock and Communication Delaysy
Time differences caused by the lack of a common time origin are
referred to as clock phase diff ( l k bi )differences or ( clock
bias)
The clocks of different nodes may disagree on h i ”0”what time”0”
means
Clock skew (drift): At real time t the computer l k i di i C( ) hi
hclock indicates time C(t), which may or may
not be the same as t.
2007/11/8 Jang-Ping Sheu, NTHU 22
Difficulties in Time SynchronizationDifficulties in Time
Synchronization
In sensor network No special master clocks are availableNo special
master clocks are available Connections are ephemeral C i ti d l i
i t t dCommunication delays are inconsistent and
unpredictable
2007/11/8 Jang-Ping Sheu, NTHU 23
The Latency in ChannelThe Latency in Channel
Send time: construct the messageSend time: construct the message
Operating system calls Context switchingg Data access to the
network interface
Access time: access transmission channel C iContention
Collisions
Propagation time: message across the channelPropagation time:
message across the channel to the destination node
Highly variable (Single hop or multi-hop)g y ( g p p) Receive time:
this time for the network interface
The delay can kept small
2007/11/8 Jang-Ping Sheu, NTHU 24
Routing ProtocolsRouting Protocols
2007/11/8 Jang-Ping Sheu, NTHU 25
Topology Control ProblemTopology Control Problem
Topology control is to design power-efficient algorithms
maintain network connectivity optimize performance metrics such as
network lifetime and throughput
Topology control determines the network topology by controlling the
transmission power of sending a physical layer broadcast
2007/11/8 Jang-Ping Sheu, NTHU 26
MotivationMotivation
No topology control: large transmission radius (with maximum
transmission radius R )(with maximum transmission radius R )
high interferencehigh interference High energy consumption Low
throughput
2007/11/8 Jang-Ping Sheu, NTHU 27
MotivationMotivation
Network mayNetwork may partition
Moti ationMotivation
Global network connectivity Li l i fLittle interference Low energy
consumptionconsumption High throughput
2007/11/8 Jang-Ping Sheu, NTHU 29
Coverage ProblemCoverage Problem
Find a subset of connected sensor nodes to cover an queried
region
Sensing node
Coverage Problemg
Each node has different sensing range and different communication
ranged e e co u ca o a ge Nodes are sufficient to cover the sensor
network and connectedand connected Finding a minimum number of
sensing nodes k- co ering a gi en region is NP hard e en k
1covering a given region is NP-hard even k = 1
2007/11/8 Jang-Ping Sheu, NTHU 31
Recently Research ResultsRecently Research Results
Octopus I – AVR (2006/3)AVR (2006/3)
ATmega128 CC2420ATmega128 CC2420
LEDs
Size: 68*25 (mm) 8-bit AVR microcontroller core @8MHz {128KB
in-system programmable flash 4KB RAM 4KB EEPROM
Size: 68*25 (mm)
4KB EEPROM ADC 10-bit 8 Channels} + IEEE 802.15.4 compliant RF
transceiver
2007/11/8 33Jang-Ping Sheu, NTHU
Size: 68mm*25mm
RS232 JTAG
connector ISP
Super Node– MSP430MSP430
Sensor
ConnectorSize: 55*30 (mm)
{40KB in-system programmable flash 10KB RAM ADC 12-Bit 8
Channels}
USB
Super Node– MSP430MSP430
Hardware FeaturesHardware Features
CC2420 - IEEE 802.15.4 wireless transceiver MSP430 - 8MHz Core
microcontrollerMSP430 8MHz Core microcontroller Expandable flash
memory – up to 1 MBytes E t l ill t F th d l dExternal oscillator –
For the deep sleep mode A connector with 50 expansion I/O pins
Sensors in advance
Two light sensors One temperature-humidity sensor
382007/11/8 Jang-Ping Sheu, NTHU
WSN Simple Node
43
2007/11/8 Jang-Ping Sheu, NTHU
C/C++ Li /Wi 32 OSLinux/Win32 OS / CCD/GPS/CCD/GPS/
1 2
1
2
1
2
1