Collecting Correlated Information from Wireless Sensor Networks

1
Collecting Correlated Information from Wireless
Sensor Networks

• Presented by Bo Han
• Joint work with Aravind Srinivasan and Amol
  Deshpande

2
Outline

• Introduction of wireless sensor networks
• Communication architecture
• Design factors of sensor networks
• Applications of sensor networks
• Correlated information collection
• What we have done so far
• Conclusion

3
Introduction

• A sensor network is composed of a large number of
  sensor nodes, which are densely deployed either
  inside the phenomenon or very close to it.
• Random deployment, the position of sensor nodes
  need not be engineered or predetermined.
• Self-organizing capabilities.
• Cooperative effort of sensor nodes.

4
Sensor Networks vs
Ad-Hoc Networks

• The number of nodes in a sensor network can be
  several orders of magnitude higher than the nodes
  in an Ad-Hoc network.
• Sensor nodes are densely deployed.
• Sensor nodes are prone to failures.
• The topology of a sensor network changes very
  frequently.

5
Sensor Networks vs
Ad-Hoc Networks

• Sensor nodes mainly use broadcast, most ad hoc
  networks are based on point-to-point
  communication.
• Sensor nodes are limited in power, computational
  capacities and memory.
• Sensor nodes may not have global ID.

6
Communication Architecture

• The sensor nodes are usually scattered in a
  sensor field.
• Each of these scattered sensor nodes has the
  capabilities to collect data and route data back
  to the sink.
• Data are routed back to the sink by a multi-hop
  infrastructureless architecture.
• The sink may communicate with the task manager
  node via Internet or satellite.

7
Example of Sensor Networks
8
Data Delivery Models

• Continuous sensors communicate their data
  continuously at a prespecified rate.
• Event driven sensors report information only
  when the event of interest occurs.
• Observer initiated (request-reply) sensors only
  reports their results in response to an explicit
  request from the observer.
• Hybrid all three approaches coexist.

9
Protocol Stack
10
Five Layers

• The physical layer addresses the needs of simple
  but robust modulation, transmission, and
  receiving techniques.
• The MAC protocol must be power-aware and able to
  minimize collision with neighbors broadcasts.
• The network layer takes care of routing the data
  supplied by the transport layer.
• The transport layer helps to maintain the flow of
  data if the sensor networks application requires
  it.
• Different types of application software can be
  built and used on the application layer.

11
Three Plans

• The power management plane manages how a sensor
  node uses its power.
• The mobility management plane detects and
  registers the movement of sensor nodes, so a
  route back to the user is always maintained, and
  the sensor nodes can keep track of who their
  neighbor sensor nodes are.
• The task management plane balances and schedules
  the sensing tasks given to a specific region.

12
Design Factors

• Fault tolerance
• Scalability
• Production costs
• Hardware constraints
• Transmission media
• Power consumption
• Sensor network topology
• Environment

13
Fault Tolerance

• Fault tolerance is the ability to sustain sensor
  network functionalities without any interruption
  due to sensor node failures.
• The fault tolerance level depends on the
  application of the sensor networks.

14
Scalability

• Depending on the application, the number may
  reach an extreme value of millions. New schemes
  must be able to work with this number of nodes.
• Basically, the density gives the number of nodes
  within the transmission radius of each node in a
  region.
• Must also utilize the high density of the sensor
  networks.

15
Production Costs

• The cost of a single node is very important to
  justify the overall cost of the networks, since
  sensor networks consist of large number of sensor
  nodes.
• The cost of a sensor node should be less than a
  dollar.

16
Transmission Media

• In a multihop sensor network, communicating nodes
  are linked by a wireless medium.
• To enable global operation, the chosen
  transmission medium must be available worldwide.
• Radio, infrared and optical media.

17
Power Consumption

• Limited power source.
• Battery lifetime is limited.
• Each sensor node plays a dual role of data
  originator and data router (data processor).
• The malfunctioning of a few nodes consumes lot of
  energy (rerouting of packets and significant
  topological changes).

18
Sensor Network Topology

• Pre-deployment and deployment phase, either
  thrown in as a mass or placed one by one.
• Post-deployment phase, topology changes are due
  to change in sensor nodes position,
  reachability, available energy, malfunctioning,
  and task details.
• Re-deployment of additional nodes phase,
  additional sensor nodes can be redeployed at any
  time to replace malfunctioning nodes or due to
  changes in task dynamics.

19
Environment

• Sensor nodes are densely deployed either very
  close or directly inside the phenomenon to be
  observed.
• They usually work unattended in remote geographic
  areas.
• They may be working in the interior of large
  machinery, at the bottom of an ocean, in a
  biologically or chemically contaminated field, in
  a battlefield beyond the enemy lines, and in a
  home or large building.

20
Applications of Sensor Networks

• Military Battlefield surveillance, Nuclear,
  biological and chemical attack detection and
  reconnaissance.
• Environment Forest fire detection, Flood
  detection.
• Health Telemonitoring of human physiological
  data, tracking and monitoring patients and
  doctors inside a hospital.
• Home application Home automation and smart
  environment.

21
Sensor Devices and Applications

• Berkeley Motes
• iBadge - UCLA
• MIT d'Arbeloff Lab The ring sensor
• Nose-on-a-chip
• Zilogs eZ80
• iButton

22
Berkeley Motes

• Small (under 1 square) microcontroller.
• It consists of
• Microprocessor
• A set of sensors for temperature, light,
  acceleration and motion
• A low power radio for communicating with other
  motes
• C compiler inclusion.

23
Berkeley Motes
24
iBadge - UCLA

• Investigate behavior of children/patient.
• Features
• Speech recording / replaying
• Position detection
• Direction detection / estimation (compass)
• Weather data temperature, humidity, pressure and
  light

25
iBadge - UCLA
26
MIT d'Arbeloff Lab The Ring Sensor

• An ambulatory, telemetric, continuous health
  monitoring device developed by d'Arbeloff
  Laboratory for Information Systems and Technology
  at MIT.
• Monitor the physiological status of the wearer
  and transmit the information to the medical
  professional over the Internet.

• Clinical trials have been done in conjunction
  with Massachusetts General Hospital's Emergency
  Room, and researchers are now working on
  commercialization of the ring-sized device.

27
Nose-on-a-chip

• Nose-on-a-chip is a MEMS-based sensor, developed
  at Oak Ridge National Laboratory.
• Can detect 400 species of gases and transmit a
  signal indicating the level to a central control
  station.

• Consists of an array of tiny sensors on one
  integrated circuit and electronics on another.
• The chip can be customized to detect virtually
  any chemical or biological species.

28
Zilogs eZ80

• Provides a way to internet-enabled process
  control and monitoring applications.
• Temperature sensor, water leak detector and many
  more applications.
• Enables users to access Webserver data and files
  from anywhere in the world.

29
iButton

• A 16mm computer chip armored in a stainless steel
  can.
• Up-to-date information can travel with a person
  or object.

30
Correlated Information Gathering

• Problem Definition
• Correlation Modeling
• Distributed Source Coding
• Asymmetric Communication Channels
• Our Approach

31
Fundamental Problem

• Collecting information from distributed sources.
• Objective correlations reduce bits that must be
  sent.
• Correlation examples in sensor networks
• Weather in geographic region.
• Similar views of same image.
• Focus information theory
• Number of bits sent.
• Ignore network topology.

• Server

32
Modeling Correlation
x1

• k sensor nodes each have n -bit string.
• Input drawn from distribution D.
• Sample specifies all kn bits.
• Captures correlations
• and a priori knowledge.
• Objective
• Inform server of all k strings.
• Ideally nodes send H(D) bits.
• H(D) Binary entropy of D.

x2
x3
x4

• Server

x5
x6
xk
33
Binary entropy of D

• Optimal code to describe sample from D
• Expected number of bits required H(D).
• Ranges from 0 to kn.
• Easy if entire sample known by single node.
• Idea shorter codewords for more likely samples.
• Challenge of this problem input distributed.

34
(Distributed) Source Coding

• Source coding is the process of encoding
  information using fewer bits than an unencoded
  representation, also called data compression.
• DSC refers to the compression of the outputs of
  two or more physically separated sources. The
  sources do not communicate with each other. These
  sources send their compressed outputs to a
  central point for joint decoding. DSC is part of
  network information theory.
• DSC has recently become a very active research
  area more than 30 years after Slepian and Wolf
  laid its theoretical foundation.

35
Distributed Source Coding

• Slepian-Wolf, 1973
• Simultaneously encode r
• independent samples.
• As r ? ?,
• Bits sent by nodes ? rH(D).
• Probability of error ? 0.
• Drawback relies on r ? ?
• Recent research try to remove this.

• Server

36
Slepian-Wolf Theorem
37
Self-Coding Foreign Coding
38
Network Coding

• Network coding is a field of information theory
  and coding theory.
• A method of attaining maximum information flow in
  a network.
• The core notion of network coding is to allow
  mixing of data at intermediate network nodes.
• A receiver sees these data packets and deduces
  from them the messages that were originally
  intended for that data sink.

39
Butterfly Network
40
New approach
x1

• Allow interactive communication!
• Nodes receive feedback from server.
• Server at least as powerful as nodes.
• Power utilization
• Central issue for sensor networks.
• Node sending is power intensive.
• Node receiving requires less power.
• .

x2
x3

• Server

x4
x5
xk
41
New approach
x1

• Communication model
• Synchronous rounds
• Nodes send bits to server.
• Server sends bits back to nodes.
• Nothing directly between nodes.
• Asymmetric Communication
• Channels
• Objectives
• Minimize bits sent by nodes.
• Ideally O(H(D)k).
• Minimize bits sent by server.
• Minimize rounds of communication.

x2
x3

• Server

x4
x5
xk
42
Who knows what?

• Server

D

• Nodes only know own string.
• Can also assume they know distribution D.
• Server knows distribution D.
• Typical in work on DSC.
• Some applications D must be learned by server
• Most such cases D varies with time.
• Crucial to have r as small as possible.

43
Fingerprint Function

• A class of functions that map an n-bit string to
  a single bit such that if f is chosen uniformly
  at random from this class, then for any y1?y2
• Prf(y1) f(y2) p, for some p ½
• An n3 x n matrix with each item chosen uniformly
  at random from 0, 1.

44
Correlation

• Correlation Multivariate Normal Distribution.
• Put sensor nodes uniformly at random in a unit
  square.
• Covariance matrix.
• Diagonal items are all 1
• Cov(x, y) d(x, y) r
• Discretization of this continuous distribution.
• Joint entropy estimation H(D).

45
Sample Set Generation

• Determine the size of sample set.
• Compute the Cholesky decomposition (matrix square
  root) of S, that is, find the unique lower
  triangular matrix A such that AAT S.
• Let Z (z1, , zn)T be a vector whose components
  are n independent standard normal variates (which
  can be generated, for example, by using the
  Box-Muller transform).
• Let X be µ AZ (here, µ 0).

46
The Algorithm

• The algorithm can figure out the sensor node
  string with high probability in log2(H(D))
  phases.
• Each phase has two rounds.
• First round sensor nodes send fingerprint bits
  to the server.
• Second round server sends feedback to the sensor
  nodes.

47
Phase i First Round

• Node sends some number of fingerprint bits to the
  server, each specified by a fingerprint function
  chosen randomly.
• When each fingerprint bit is processed, the
  inputs in the sample set that do not agree with
  that bit are discarded from consideration.
• Before processing each fingerprint bit sent by
  the node, the server checks to see if it is an
  unbalanced bit if some string x agrees with more
  than half of the elements of the sample set.
• The string is called heavy string.

48
Phase i Second Round

• The server sends the heavy string to the node.
• Node responses with a single indicator bit.
• The server keeps track of the total number of
  balanced bits that it has received.
• The server continues this process until it has
  received some predetermined number of balanced
  bits or the sample set is empty.

49
Conclusions

• Lots of interesting problems in wireless sensor
  networks.
• New technique for collecting distributed and
  correlated information (ongoing project).
• Allows for arbitrary distributions and
  correlations.
• Single sample sufficient.
• Open problems
• Lower bound on rounds.
• Incorporating network topology.

50
Reference

• Top conference MobiCom, MobiHoc, Sensys, IPSN,
  Infocom, SECON, SODA, STOC, FOCS.
• I.F. Akyildiz, W. Su, Y. Sankarasubramaniam, and
  E. Cayirci, Wireless Sensor Networks a Survey,
  Computer Networks 38 (2002) 393422.
• Z. Xiong, A. Liveris, and S. Cheng, Distributed
  Source Coding for Sensor Networks, IEEE Signal
  Processing Magazine 21 (2004) 8094.
• A. Ramamoorthy, K. Jain, P. Chou, and M. Effros,
  Separating Distributed Source Coding from Network
  Coding, IEEE/ACM Transactions on Networking 14
  (2006) 27852795.
• J. Liu, M. Adler, D. Towsley, and C. Zhang, On
  Optimal Communication Cost for Gathering
  Correlated Data through Wireless Sensor Networks,
  In Proceedings of MobiCom 2006.
• T. Batu, S. Dasgupta, R. Kumar, and R. Rubinfeld,
  The Complexity of Approximating Entropy, In
  Proceedings of STOC 2002.
• M. Adler, Collecting Correlated Information from
  a Sensor Network, In Proceedings of SODA 2005.

51
Questions?