Dynamic Sensor Networks
Event
Smart
Sensor
Node
COTS
PDA
Smart
Sensor
Node
Target
Target
Smart
Sensor
Node
Impact
 GPS leveraged for geo-referenced identity, and low
power communications synchronization. Up to 100x
communications power reduction.
 Standard APIs implemented as Java class libraries and
browser-based user interfaces provide code mobility,
code reuse, and platform independence.
 High-level spatial and context “anycast” addressing
enables dynamic specialization for augmented
awareness and collaborative consensus applications.
New Ideas
 Power-aware link and routing protocols. Exploit finegrained power control of radios for energy efficient
connectivity. Maximize sensor network’s operational
lifetime through energy-aware routing.
 GPS-aware link protocols. GPS-synchronized
ultra-low-power communication.
 Spatial addressing and connectivity. High-level
addressing, unicast, multicast, anycast, and gathercast
communication based on spatial referencing of the nodes.
 Mobile code and web technology. Embedded Java APIs
for code portability and browser-based topographical map
interface for visualizing dynamic data from sensor net.
Milestones
Sensor Control API Specification
Topographical Map Interface Definition
Network Services API Specification
GPS-Aware Link Protocol ExperimentFY01 Q4
Network Services PDA/Laptop Experiment
Integrated Sensor-Kit Experiment
Brian Schott PI, Bob Parker (USC/ISI), Mani Srivastava (UCLA) Co-PI, Mark Jones (Virginia Tech) Co-PI
FY00 Q1
FY00 Q1
FY00 Q2
FY01 Q4
FY02 Q4
Dynamic Sensor Networks
• DSN is focusing on three
SenseIT areas:
1) Platforms
– GPS-synchronized ultra-lowpower communication
experimental platform.
2) Distribution and Aggregation
Smart
Sensor
Node
COTS
PDA
Smart
Sensor
Node
Target
Target
Smart
Sensor
Node
– Network boot-up, low-power
link protocols, power-aware
routing, and spatial addressing.
3) Declarative Language and
Execution Environment
– Topographical map interface.
– Java APIs for portability.
– sensor network emulation for
rapid application development.
DSN Experimental Platform
DSN Experimental Platform
• The primary purpose of the platform is to experiment with using GPS
synchronization to provide precise control over transmit/receive on radio.
• Uses COTS hardware/software to minimize cost and maximize code
portability. No attempt to miniaturize subsystems.
Comm
Subsystem
Serial port
GPS
Antenna
GPS
Radio
Serial port / TOD
Comm
Antenna
Sensor
Subsystem
Sensors
PDA
Comm Subsystem
Sensor Subsystem
Application specific
processing done in COTS
PDA. Assumed in sleep
mode unless activated by
Comm Subsystem. May
have other high-power
sensors in PCMCIA slot.
Synchronizes radios
using GPS signal. Can
store/forward packets in
sensor net and perform
routing without PDA.
Goal: Runs on solar cell.
Contains sensors, signal
conditioning, signal
processing, and
optionally protocol
processing to eliminate
PDA for light-weight
sensor node.
DSN Comm Subsystem (Concept)
Clock
Oscillator
Sleep
Mode
Variable Ratio
Divider
TOD to Sensor
Subsystem
TOD
(Time Of Day)
Data
Timing
CLK IN
Serial Port
Controller
To
PDA
GPS
Antenna
To
Sensor
Subsystem
External Devices
(i.e.. Long-range radio)
Data
Buffer
Microcontroller
GPS
Radio
1 pulse per second
Comm
Antenna
Distribution and Aggregation
Mani Srivastava
UCLA (Co-PI)
Distribution and Aggregation
•
•
•
Distribution of node location and
capabilities to neighborhood
application query servers at bootup and reconfiguration.
Low-power link protocols and
power-aware routing for energy
efficient sensor data distribution.
Global spatial addressing that
support referencing of individual or
groups of sensor nodes by
geographic location
Capability-based addressing in the
local neighborhood.
External IP connectivity with DSN
network gateways.
•
•
+
S
+
S
-
S
Target
S
-
+
S
S
+
S
S
+
-
+
S
Sensor Network
-
-
Internet
Power-Aware Link Protocols
• Optimize for computation and communication energy
spent per bit distributed (Joules/bit) as opposed to
traditional metrics such as throughput.
• Combine novel channel state estimation techniques with
the capabilities of radios to adapt transmit power and other
parameters such as spreading gain and symbol rate.
• Exploit GPS reference timing signal
to synchronize communication among
sensor nodes and minimize
transmit/receive windows.
S
S
S
S
S
Example of Impact of Link Layer
Adaptation on Energy Efficiency
14
12
Energy per useful bit (J/bit)
10
BER=10-4
BER=10-3
8
6
BER=10-8
4
2
0
200
400
600
800
1000
Packet Length (bytes)
1200
1400
Power-Aware Routing Protocols
• Traditional multihop ad hoc
routing protocols focus on fast
topology changes.
• Novel DSN routing protocols
will focus on maximizing
sensor network lifetime.
– Power hot-spots that lead to
holes in coverage and network
partitioning.
– Power inefficiency due to
signaling messages in
quiescent state.
– Power-based routing metrics.
– Leveraging location
information during routing.
+
S
+
+
S
S
+
S
-
-
+
S
+
S
-
+
S
-
-
-
S
Target
Spatial Addressing
• Sensor applications typically not interested in node IDs.
– Query destination in terms of node location and capabilities.
– Any suitable node in target neighborhood can handle the query.
• DSN network addressing architecture and routing
protocols optimized for the needs of sensor applications.
–
–
–
–
–
Node addresses encode their location.
Location-based global routing and capability-based local routing.
Intermediate nodes intelligently filter/combine query responses.
Java-based API for communicating with nodes and node groups.
Both native DSN architecture as well as IPv6 overlay-based will
be investigated for use within the sensor network.
• External users can tunnel into network across the IPv6
Internet via gateway nodes.
DSN Network Services API
• Network Services Library
provides multihop connectivity
and higher-level networking
services.
• Java API choice provides
portability to other wired and
wireless platforms.
• Low-level link layer accessible
to other efforts implementing
their own routing protocols.
• API also supports DSN
emulation environment.
Applications
Java-based API
Run-time Environment
Spatial addressing
Power-aware spatial routing
Low-power link protocols
Sensor Node Hardware
Sensor
Network
Emulation
ns-based DSN Simulation
• ns will be used as the primary DSN simulation platform.
– Simulation of link and routing protocols.
– Comparison of alternative addressing architectures.
• Collaboration with Deborah Estrin’s SCADDS to define
common ns modules and interfaces.
– Preliminary interaction already started.
• ns module for DSN protocol stack will be made available
to other members of SenseIT community.
Declarative Languages
and Execution Environment
User Interface
• PDA-based platform
–
–
–
–
Windows CE
Java 2 (Micro Edition)
Color screen
Pen-based interface
• Connects to DSN Comm
Subsystem by serial port.
• Interface will easily operate on
more powerful Java platforms.
Topographical Interface
• Topographical map interface to
provide network status visually.
• Queries can be geographic, by
sensor id/capabilities, or be
application specific.
• Input using pop-up forms/menus.
• Browser plug-in approach to
allow multiple applications to run
on the same map.
• User query translated to query
language defined by other
SenseIT effort(s).
Any tracked vehicles in
this region?
Sensor Network Query Action
•
•
•
•
User submits query via map browser interface.
Local application (one of several) interprets query.
Query sent using query language over network.
Query arrives at appropriate application query server based
on geographic address.
• Query is sent to
appropriate sensor
nodes.
• Local Java
application
acts on query.
Java Emulation Environment
• Java-based sensor network
emulation tools for application
development.
– Allows for testing of components
before all hardware is ready.
– Allows for testing on a larger number
of sensors nodes than available.
• The goal is to enable rapid
application development by
providing source-code
compatibility between
a workstation emulator
and sensor node.
Summary
New Ideas
 Power-aware link and routing protocols. Exploit fine-grained power control of radios
for energy efficient connectivity. Maximize sensor network’s operational lifetime
through energy-aware routing.
 GPS-aware link protocols. GPS-synchronized ultra-low-power communication.
 Spatial addressing and connectivity. High-level addressing, unicast, multicast, anycast,
and gathercast communication based on spatial referencing of the nodes.
 Mobile code and web technology. Embedded Java APIs for code portability and
browser-based topographical interface for visualizing dynamic data from sensor net.
Impact
 GPS leveraged for geo-referenced identity, and low power communications
synchronization. Up to 100x communications power reduction.
 Standard APIs implemented as Java class libraries and browser-based user interfaces
provide code mobility, code reuse, and platform independence.
 High-level spatial and context “anycast” addressing enables dynamic specialization for
augmented awareness and collaborative consensus applications
Descargar

Dynamic Sensor Networks - University of North Texas