Transcript Discovering Sensor Networks - IEEE Real World Engineering Projects

Discovering Sensor Networks:
Applications in Structural Health
Monitoring
Summary Lecture: Part 4
Wireless Sensor Networks
Discovering Sensor Networks:
Applications in Structural Health Monitoring
• Distributed networks of wireless
sensor nodes
– gather critical information about the
physical world
– communicate the information to
remotely located decision makers
• Example applications:
– smart farming
– health applications for the elderly
– environmental monitoring
– monitoring the structural health of a
building or bridge
Our Original Motivation: Structural Health
Monitoring for Disaster Prevention
bridges
dams
public venues
We Learned about…
• Key concepts in Electrical and Computer Engineering, in
particular:
1. Sensor read-out electronics and data conversion
2. Introduction to MEMS, microsystems and sensors
3. Radio-frequency (RF) wireless data communications
4. Wireless sensor networks
• Practical steps in the implementation of a wireless
sensor network to monitor structural health
– From a sensed parameter, through sensor readout
and data conversion, to wireless transmission and
forming a network
Transducers
(ELECTRONICS)
Sensor
Readout
ACTUATORS
Analog
CONTROL PARAMETERS
Communications
with Wireless
Network
SIGNAL
CONDITIONING
Power &
Control
SIGNAL
CONDITIONING
(ELECTRONICS)
Digital
MICROPROCESSOR
SENSORS
SENSED PARAMETERS
Wireless Sensor Systems
RF
RF/Wireless
Transceiver
The Big Picture: Our Sensor Network Story
1. A physical phenomenon (deflection, in our case) is
converted into an electrical parameter (resistance).
• Piezoresistive strain gages:
metallic films whose electrical
resistance varies with the strain
• We also observed that the
performance of electronic
devices and circuits can have
undesired variation in their
performance due to
temperature, mechanical
stress, light, etc.
The Big Picture: Our Sensor Network Story
2. The electrical parameter is converted to an analog
signal via readout circuitry.
• The Wheatstone bridge
• Assuming the bridge starts
off in the balanced condition,
if Rx varies due to an
environmental condition (e.g.
temperature, stress, etc.), a
non-zero voltage will be
detected across nodes D and
B
The Big Picture: Our Sensor Network Story
3. Analog-to-digital conversion.
ADC
101 110 100 011 100 110 111…
• Note that the output of the DAC is not a perfect
representation of the original analog signal → we call
this quantization error
The Big Picture: Our Sensor Network Story
4. Digital data stored in microprocessor or memory.
• The bridge output voltage data
is converted to digital bits by a
Freescale MC9S12C32 16-bit
microcontroller unit (MCU) with
on-board analog-to-digital
converter (ADC) and
transferred to students’ laptops
via USB cable
• The project board DC supply
rails are powered via the USB
cable from a student’s laptop
The Big Picture: Our Sensor Network Story
5. Data formatted/encoded for transmission.
• Transmission uses Freescale
AP13192USLK ZigBee
transceivers
• ZigBee devices conform to the
IEEE 802.15.4-2003 Low-Rate
Wireless Personal Area Network
(WPAN) standard
• This standard specifies operation
in the unlicensed 2.4 GHz,
915 MHz and 868 MHz Industrial /
Scientific / Medical (ISM) bands
The Big Picture: Our Sensor Network Story
6. Formatted/coded data modulated on RF carrier
(Zigbee uses something called QPSK, quadrature phase
shift keying) and transmitted over the wireless channel.
DAC
101 110 100 011 100 110 111…
×
The Big Picture: Our Sensor Network Story
7. Use of the channel by network nodes is determined
by the Medium Access Control protocol.
•
•
•
•
Aloha
Simple random access
If have data to send, just
send it
If collision occurs, try to
resend again later
Tends to be inefficient
CSMA
• Random access, not as
simple as Aloha
• If have data to send, listen
first:
• channel free? send
• channel occupied? listen
again later
• If collision occurs, try again
later
The Big Picture: Our Sensor Network Story
8. Waveform received at the other node is demodulated
and converted back to digital.
101 110 100 011 100 110 111…
The Big Picture: Our Sensor Network Story
9. Network of nodes allows for sensing along entire
bridge span.
• Sensors may form a multi-hop
wireless network
– Or they may all report directly
to a sink
• Information is synthesized and
analyzed at the sink
– Which, in turn, reports critical
situations to some commandand-control center
Annotated map of
sensor field:
Average:
sensor
Mission Accomplished: Disaster Averted
• During lab:
• Post-lab:
– Sensor node underwent
– Convert received voltages
deflection: a digital
to resistances, R
voltage resulted
– Using relationship between
– Voltage transmitted
R, G and strain (e),
throughout network
compute strain levels at
each sensor node
– Received by other sensor
nodes (if in range and no
– Compare to strain
collisions)
threshold to determine
unsafe strain levels
– Generate warning and alert
safety personnel, result
Think / Pair / Share
• Consider the wireless sensor system block diagram.
• What sources of error can you identify that will result in
the output data at the receiving node deviating from the
actual physical parameter being sensed?
Think / Pair / Share – Possible Answers
• What sources of error can you identify that will result in
the output data at the receiving node deviating from the
actual physical parameter being sensed?
– Error in conversion of physical parameter to
electrical parameter (readout circuit nonlinearity,
calibration error…)
– Quantization error in sensor ADC
– Bit error in RF transmitter
– Channel impairments (propagation loss, fading,
noise, interference…)
– Data packet collisions
– Bit error in RF receiver
– Quantization error in receiver ADC
Concepts, Trade-offs, and Tools
Concepts Discovered




Perform analog-todigital data
conversion
Sense mechanical
movement and
convert it to
electrical data
Transmit data
efficiently through
a wireless medium
Establish a
wireless sensor
network and
transmit/receive
data between the
sensor nodes
Solution Strategies




Balanced
Wheatstone bridge
circuit for sensing
resistance
changes
Sensor to convert
beam deflection to
electrical
resistance
Carrier sensing to
improve data
throughput
Self-organizing
wireless network
to relay and
aggregate sensed
data
Trade-offs





Accuracy vs.

quantization error
Sensor accuracy
vs. speed
MAC scheme vs.
efficiency
Sensor location

vs. successful
data reception
throughout sensor 
network
Size/energy
consumption of
sensor node vs.
performance
Tools
Freescale
protoboard with
readout circuit,
microprocessor,
and Zigbee
transceiver
Strain gauge
sensor with
attached “beam”
CodeWarrior
software interface
Other Applications
lake water quality monitoring
precision agriculture
motion analysis
A Final Observation
• The development of sensors and wireless sensor
networks is highly interdisciplinary, requiring team
members with expertise spanning multiple areas of ECE
and CS:
– Sensor devices (materials science, microelectronics,
MEMS)
– Circuits and electronics
– Digital signal processing
– RF/wireless communications
– Radiowave propagation (electromagnetics)
– Wireless networking
– Software (control, embedded computing, etc.)