WIRELESS

1
WIRELESS INTEGRATED SENSOR NETWORKS (WINS)
2
WHAT DOES WINS PROVIDE?

• Distributed Network And Internet Access To
  Sensors
• Applications
• Transportation
• Health Care
• Manufacturing
• Safety and Security
• Processing At The Source

3
MAIN POINTS IN THE PAPER

• Physical principles leading to dense sensor
  networks
• Why distributed and layered signal processing??
• How to embed WINS nodes in the Internet??
• Remote Internet control and analysis of
  sensor-network operation.

4
PHYSICAL PRINCIPLES

• Propagation laws for Sensing
• All Signals decay with distance.
• Many obstructions can render electromagnetic
  sensors useless
• If the system is to detect objects reliably, it
  has to be distributed.

5
DETECTION AND ESTIMATION THEORY FUNDAMENTALS

• A Set of observables Xj and a set of hypothesis
  hi
• The set of parameters known as feature set fk
• Constructing Decision space to choose among
  hypothesis
• As feature space ? -gt complexity ?
• Hence Dense Sensor Networks
• When target is nearby decision space ?

6
COMMUNICATIONS CONSTRAINTS

• Spatial separation is very important factor
• The losses make long-range communication a
  power-hungry exercise.
• Multipath propagation is a serious problem
• It can be remedied by spreading of signal in the
  domain of frequency
• Wavefront obstruction can be dealt with multihop
  network
• This diversity reduces the power required to
  transmit data.

7
ENERGY CONSUMPTION IN INTEGRATED CIRCUITS

• Overall systems costs are large if energy system
  is large
• CMOS transistor pair draws power each time it is
  flipped
• When transmitting and receiving the signal, large
  power is consumed
• Hence, Transceiver should be off as much of time
  as possible
• Application Specific Integrated Circuits (ASICs)
  consume less energy than DSPs

8
SIGNAL PROCESSING ARCHITECTURES

• Much less power is consumed in processing data
  than transmitting it
• Two design principles to achieve reliable
  decisions with low power
• Processing hierarchy is needed
• Processing hierarchy is intertwined with
  networking and storage issues
• Given certain parameters, energy cost of
  transmitting 1kb over 100mtrs is 3 joules whereas
  processing with same energy executes 3 million
  instructions.

9

• Collaborative processing can extend the effective
  range of sensors

10
WINS NETWORK ARCHITECTURE

• Takes advantage of short distance separation and
  low power to provide multihop communication
• Devices MAC address in a network should include
  some variant of TDMA
• A time division protocol requires that
  transceivers often exchange short messages.

11
EMBEDDING WINS NODES IN THE INTERNET

• Wide range of applications
• Needs WINS gateways to provide support to WINS
  network
• WINS system design gives following advantages
• Reduced operating power, improved bit rate and
  bit error rate, simplified protocol and reduced
  cost.
• Issues of employing TCP/IPv6 in WINS

12
(No Transcript)
13
WINS NODE ARCHITECURE

• History of WINS (How WINS evolved?)
• Its development initiated in 1993 at UCLA. Later
  DARPAs LWIM project demonstrated similar
  architecture
• WINS NG architecture was developed by Sensor.com
• Future technologies of WINS nodes.

14
Critique

• Main Contribution
• Proposed new architecture as a distributed sensor
  application
• Embed WINS nodes in Internet
• Claims
• If the system has to detect objects reliably,
  then it has to be distributed
• It pays to process data locally to reduce traffic
  volume.

15

• Assumptions
• The best possible alternative for sensor network
  is processing close to sensing
• Saving energy and reducing traffic volume using
  multi hop ad hoc routing and advanced
  communication techniques
• Unstated Facts
• Control model of architecture not clearly defined
• Communication issues with the Internet still
  unresolved

16
Summary

• WINS combines microsensor technology and
  low-power signal processing and low cost
  networking into compact system.
• It describes that if reliability is needed then
  distributed system is essential.
• Embedding nodes into internet can be a
  revolutionary feature