int598 sensor networks silvia nittel spatial information science & engineering university of...
TRANSCRIPT
INT598Sensor Networks
Silvia NittelSpatial Information Science & Engineering
University of MaineFall 2006
INT598: IGERT in Sensor Science, Engineering and Informatics 2
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
IGERT
INT598: IGERT in Sensor Science, Engineering and Informatics 3
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Overview
Motivation & Applications Platforms, Operating Systems,
Power Networking
Protocols, naming, routing Data Collection and Aggregation
INT598: IGERT in Sensor Science, Engineering and Informatics 4
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Motivation
Trends: Developments of new sensor materials Miniaturization of microelectronics Wireless communication
Consequences: Embedding devices into almost any man-made and
some natural devices, and connecting the device to an infinite network of other
devices, to perform tasks, without human intervention. Information technology becomes omnipresent.
”Pervasive Computing”: The idea that technology is to move beyond the personal computer to everyday devices with embedded technology and connectivity as computing devices become progressively smaller and more powerful.
INT598: IGERT in Sensor Science, Engineering and Informatics 5
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Embedded Networked Sensing Potential• Micro-sensors, on-
board processing, and wireless interfaces all feasible at very small scale– can monitor
phenomena “up close” in non-intrusive way
• Will enable spatially and temporally dense environmental monitoring
• Embedded & Networked Sensing will reveal previously unobservable phenomena
Habitat Monitoring
Storm petrels on Maine’s Great Duck Island
Contaminant Transport
Marine Microorganisms Vehicle Detection
INT598: IGERT in Sensor Science, Engineering and Informatics 6
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Multiscale Observation and Fusion: Example, Regional (or greater) scale to local scale
images from Susan Ustin, UC Davis
Satellite, airborne remote sensing data sets at regular time intervals
coupled to regional-scale “backbone” sensor network for ground-based observations
fusion, interpolation tools based on large-scale computational models
Small-scaleSensor network
INT598: IGERT in Sensor Science, Engineering and Informatics 7
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Overview
Motivation & Applications Platforms, Operating Systems,
Power Networking
Protocols, naming, routing Data Collection and Aggregation
In-network data aggregation
INT598: IGERT in Sensor Science, Engineering and Informatics 8
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Emergence of WiSeNets 1994 Pottie and Kaiser propose Low Power Wireless Integrated
Microsensors DARPA Sensit Program (Sensor Information Technology) Late 97-98 handhelds emerge
Palm platform ITSY, BWRC PicoRadio, etc. Matchbox PCs Bluetooth promised
Berkeley SmartDust 1999 WeC mote offshoot
2000 Mote/TinyOS platforms WINS finally appears in Linux for Darpa’s Sensit 2002 Mica NEST OEP creates de facto platform 2003 Bluetooth revival 2004 Telos, lowest power mote, supports IEEE 802.15.4
INT598: IGERT in Sensor Science, Engineering and Informatics 9
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Abbreviations
Sensit Darpa’s Program “Sensor Information
Technology” WINS
Wireless Integrated Network Sensor Platforms Developed by Sensoria Corporation for Darpa’s
Sensit program NEST
Network Embedded Systems OEP
Open Experimental Platform (a middleware for sensor networks)
INT598: IGERT in Sensor Science, Engineering and Informatics 10
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Sensor Network
• “Sensor Node”:• Tiny vanilla computer with operating system, on-
board sensor(s) and wireless communication (“PC on a pin tip”)
• Trend towards low-cost, micro-sized sensors• Use of wireless low range RF communication• Batteries as energy resource
• “Sensor Network”• Massive numbers of “sensors” in the environment
that measure and monitor physical phenomena • Local interaction and collaboration of sensors• Global monitoring• Tightly coupled to the physical world to sense and
influence it
INT598: IGERT in Sensor Science, Engineering and Informatics 11
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
UC Berkeley Family of Motes
INT598: IGERT in Sensor Science, Engineering and Informatics 12
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Mica2 and Mica2Dot Processor:
ATmega128 CPU RAM/Storage:
Chipcon CC1000 Manchester
encoding Tunable frequency Byte spooling
Power usage scales with range
1 inch
INT598: IGERT in Sensor Science, Engineering and Informatics 13
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Basic Sensor Board
Light (Photo) Temperature Prototyping
space for new hardware designs
INT598: IGERT in Sensor Science, Engineering and Informatics 14
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Mica Sensor Board Light (Photo) Temperature Acceleration
2 axis Resolution:
±2mg Magnetometer
Resolution: 134G
Microphone Tone Detector Sounder
4.5kHz
INT598: IGERT in Sensor Science, Engineering and Informatics 15
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Mica Weather Board
Total Solar Radiation Photosynthetically
Active Radiation Resolution: 0.3A/W
Relative Humidity Accuracy: ±2%
Barometric Pressure Accuracy: ±1.5mbar
Temperature Accuracy: ±0.01oC
Acceleration 2 axis Resolution: ±2mg
Designed by UCB w/ Crossbow and UCLA
Revision 1.5
Revision 1.0
INT598: IGERT in Sensor Science, Engineering and Informatics 16
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Telos: New OEP Mote Single board philosophy
Robustness, Ease of use, Lower Cost Integrated Humidity & Temperature sensor
First platform to use 802.15.4 CC2420 radio, 2.4 GHz, 250 kbps (12x mica2) 3x RX power consumption of CC1000, 1/3 turn on time Same TX power as CC1000
Motorola HCS08 processor Lower power consumption, 1.8V operation,
faster wakeup time 40 MHz CPU clock, 4K RAM
Package Integrated onboard antenna +3dBi gain Removed 51-pin connector Everything USB & Ethernet based 2/3 A or 2 AA batteries Weatherproof packaging
Support in upcoming TinyOS 1.1.3 Release Co-designed by UC Berkeley and Intel Research Available from Moteiv (moteiv.com)
INT598: IGERT in Sensor Science, Engineering and Informatics 17
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
COTS-BOTS (UCB)Commercial Off-The-Shelf roBOTS
5” x 2.5” x 3” size <$250 total 2-axis
accelerometer
INT598: IGERT in Sensor Science, Engineering and Informatics 18
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Robomote (USC) Less than 0.000047m3
$150 each Platform to test algorithms for adaptive wireless networks
with autonomous robots
INT598: IGERT in Sensor Science, Engineering and Informatics 19
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
A Network
S. Madden, UBerkeley
INT598: IGERT in Sensor Science, Engineering and Informatics 20
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Wireless Sensor Networks
They present a range of computer systems challenges because they are closely coupled to the physical world with all its unpredictable variation, noise, and
asynchrony; they involve many energy-constrained,
resource-limited devices operating in concert;
they must be largely self-organizing and self-maintaining; and
they must be robust despite significant noise, loss, and failure.
INT598: IGERT in Sensor Science, Engineering and Informatics 21
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Sensor Network Objectives Several classes of systems:
Mote herds Collaborative processing
arrays (32 bit, 802.11, linux) Networked Info-Mechanical
Systems: Autonomy Achieve longevity/autonomy,
scalability, performance with: heterogeneous systems in-network processing,
triggering, actuation Algorithm/Software challenges
Characterizing sensing uncertainty
Error resiliency, integrity Statistical and information-
theoretic foundations for adaptive sampling, fusion
Programming abstractions, Common services, tools
Data modeling, informatics
lifetime/autonomy
scale
Collaborative processing arrays (imaging, acoustics)
samplingrate
Mote Clusters
INT598: IGERT in Sensor Science, Engineering and Informatics 22
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Sensor Network Design Topics
Long-lived systems that can be untethered (wireless) and unattended
Communication will be the persistent primary consumer of scarce energy resources (MICA Mote: 720nJ/bit xmit, 4nJ/op)
Autonomy requires robust, adaptive, self-configuring systems
Leverage data processing inside the network Exploit computation near data to reduce communication, achieve
scalability Collaborative signal processing Achieve desired global behavior with localized algorithms
(distributed control)
“The network is the sensor” (Manges&Smith, Oakridge Natl Labs, 10/98)
Requires robust distributed systems of hundreds of physically-embedded, unattended, and often untethered, devices.
INT598: IGERT in Sensor Science, Engineering and Informatics 23
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Architecture
Data aggregation, Query processing
Adaptive topology, Geo-Routing
MAC, time, location
Phy: comm, sensing, actuation
Data model, Declarative queries
Application: Events, Reactions
Network layer
(temp-spatial)DB layer
Physical layer
Application layer
Source: Deborah Estrin, UCLA
INT598: IGERT in Sensor Science, Engineering and Informatics 24
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Overview
Motivation & Applications Platforms, Operating Systems,
Power Networking
Protocols, naming, routing Data Collection and Aggregation
INT598: IGERT in Sensor Science, Engineering and Informatics 25
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Communication using Radio
Broadcastingradio signals
Listening &receiving signals
INT598: IGERT in Sensor Science, Engineering and Informatics 26
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Energy required to transmit signals in distance d Communication is huge battery drain Indoor has lots of other complications
Small energy consumption => short range comm Multihop routing required to achieve distance Routes around obstacles Requires discovery, network topology formation,
maintenance may dominate cost of communication
Energy to receive ~ E*t at short range Dominated by listening time (potential receive) Radio must be OFF most of the time!
PicoRadio and Radio propagation
INT598: IGERT in Sensor Science, Engineering and Informatics 27
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
ISO/OSI Protocol Stack
PhysicalData LinkNetwork
TransportSession
PresentationApplication
7 Layer ISO/OSI Reference Model
The NetworkCard
The InternetProtocols
InternetApplication
The End Computer System View
Transport Control Protocol (TCP)
Internet Protocol
(IP)
*) International Standard Organization's Open System Interconnect
INT598: IGERT in Sensor Science, Engineering and Informatics 28
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Low-level Networking Physical Layer
Low-range radio broadcast/receive Wireless (wiSeNets)
MAC: Media Access Control Controls when and how each node can transmit in the wireless
channel (“Admission control”) Objectives:
Channel utilization How well is the channel used? (bandwidth utilization)
Latency Delay from sender to receiver; single hop or multi-hop
Throughput Amount of data transferred from sender to receiver per time unit
Fairness Can nodes share the channel equally?
INT598: IGERT in Sensor Science, Engineering and Informatics 29
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
MAC Design Decisions
Energy is primary concern in sensor networks
What causes energy waste? Collisions Control packet overhead Overhearing unnecessary traffic Long idle time
bursty traffic in sensor-net apps Idle listening consumes 50—100% of the power
for receiving (Stemm97, Kasten)
Dominant factor
INT598: IGERT in Sensor Science, Engineering and Informatics 30
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Networking Network Architecture: Can we adapt the Internet
protocols and the “end to end” architecture to SN? Internet routes data using IP Addresses in Packets
and Lookup tables in routers Many levels of indirection between data name and IP
address, but basically address-oriented routing Works well for the Internet, and for support of Person-to-
Person communication
Embedded, energy-constrained (un-tethered, small-form-factor), unattended systems cannot tolerate communication overhead of indirection
Our sensor network architecture needs Minimal overhead Data centric routing
INT598: IGERT in Sensor Science, Engineering and Informatics 31
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Data-centric Routing Named-data as a way of tasking motes, expressing data
transport request (data-centric routing) Basically:
“send the request to sensors that can deliver the data, I do not care about their address”
Two initial approaches in literature: Derived from multicast-routing perspective
where you name a logical group of sensor nodes (Diffusion)
Derived from database query language (TinyDB) with stronger semantics on data delivery, timing, sequencing
Commonality is tree-based routing Query sent out from microserver to motes Sink-Tree built to carry data from motes to
microserver
INT598: IGERT in Sensor Science, Engineering and Informatics 32
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Tree Routing
A
B C
D
FE
Query
Parent Node
Children Nodes
INT598: IGERT in Sensor Science, Engineering and Informatics 33
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Tree building Queries/Request
What goes in query? Where does query go?
Neighbor selection How does mote select upstream neighbor for
data? Asymmetric links Unidirectional links Route characterization (like ETX)
Multiple microservers What about multiple microservers? How does mote select a microserver?
INT598: IGERT in Sensor Science, Engineering and Informatics 34
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Tree building
Dynamics How often do you send out a new query? How often do you select a new upstream path
Design Tree building protocol From query source to data producer(s) and back Multihop ad-hoc routing
reliable routing is essential!
INT598: IGERT in Sensor Science, Engineering and Informatics 35
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Basic Primitives Single Hop packet loss characteristics
Environment, distance, transmit power, temporal correlation, data rate, packet size
Services for High Level Protocols/Applications Link estimation Neighborhood management Reliable multihop routing for data collection
INT598: IGERT in Sensor Science, Engineering and Informatics 36
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Basic Neighborhood of Devices
Services for High Level Protocols/Applications
Link estimation Neighborhood management Reliable multihop routing for data
collection
Direct Reception Large variation in affinity
Asymmetric links Long, stable high quality links Short bad ones
Link quality varies with traffic load Collisions Distant nodes raise noise floor
Many poor “neighbors”
INT598: IGERT in Sensor Science, Engineering and Informatics 37
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Neighborhood Management Maintain link estimation statistics and routing
information of each neighboring sensor node How large should this table be?
O(cell density) * meta-data for each neighbor Issue:
Density of nodes can be high but memory of each node is limited
At high density, many links are poor or asymmetric Neighborhood Management
Question: when table becomes full, should we add new neighbor? If so, evict which old neighbor?
Similar to frequency estimation of data streams, or classical cache policy
INT598: IGERT in Sensor Science, Engineering and Informatics 38
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Reliable Routing 3 core components for Routing
Neighbor table management Link estimation Routing protocol
INT598: IGERT in Sensor Science, Engineering and Informatics 39
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Routing Protocols
Ad-hoc routing, Geographic routing Topology Formation Directed Diffusion Rumor/Gossip Routing
INT598: IGERT in Sensor Science, Engineering and Informatics 40
© Dr. Silvia Nittel, NCGIA, University of Maine, 2006
Overview
Motivation & Applications Platforms, Operating Systems,
Power Networking
Physical layer, MAC, Protocols Routing
Adaptable, Configurable Systems Data Collection and Aggregation