CS 851 Wireless Sensor Networks Introductory Lecture

1
CS 851Wireless Sensor NetworksIntroductory
Lecture
Professor Jack Stankovic Department of Computer
Science University of Virginia September 2003
2
Purpose of this Lecture

• Get you to think differently
• Regardless of whether you are new to WSN or have
  been working with them
• Introduce the basic key issues and their
  implications
• Reduce work to its essence

3
The field is exploding
4
Smart Spaces
Smart School
Smart Factory
Smart City

• Other Applications
• Battlefields/Surveillance
• Earthquake areas
• Environmental Monitoring
• Airport security
• Emergency Response
• Location Services

5
More Applications

• Interface with the Internet
• Handheld PDAs/laptops
• Element in pervasive computing

From your reading did you find interesting applica
tions or ideas about applications that
were Surprising?
6
Ad Hoc Wireless Sensor Networks

• Sensors
• Actuators
• CPUs/Memory
• Radio

7
Research Questions

• What are the correct HW elements to make
  solutions at the OS/middleware/application levels
  easier?
• Current motes are only 1 possible platform
• How about DSPs? Special security HW?
• What capacities (cpu speed, memory, bandwidth,
  power, etc.) and their fundamental limitations,
  have if any, on solutions

8
Sensor/Actuator Clouds
Resource management, team formation, networking,

Heterogeneous Homogeneous

Severe constraints power, memory, bandwidth,
cpu, cost, ...
9

• Background Challenge fundamental assumptions
  underlying distributed systems technology
• How the problems change
• Key Areas to be Addressed
• Routing
• Power Management
• Localization
• Security
• Paradigms
• Theory
• Other Issues
• Examples key research problems/solutions
• Spatial-Temporal Routing
• Application Independent Data Aggregation
• Localization Realities

10
How the Problems Change

• Environment
• connect to physical environment (large numbers,
  dense, real-time)
• massively parallel interfaces (sometimes)
• faulty, highly dynamic, non-deterministic
• wireless (indirect impact on remote entity)
• power management critical
• Network
• structure is dynamically changing
• sporadic connectivity
• new resources entering/leaving
• large amounts of redundancy
• self-configure/re-configure
• individual nodes are unimportant - route/query to
  AREA

11
How the Problems Change

• OS/Middleware
• manage aggregate performance
• control the system to achieve required emerging
  behavior
• How do we know it works?
• self-organizing (self-)
• fuzzy membership and team formation
• manage power/mobility/real-time/security
  tradeoffs
• geographical/location based (spatial)
• real-time/real world (temporal)
• data centric

12
Examples

• Can you give me examples of simple decentralized
  algorithms that exhibit aggregate behavior?

13
Implications

• Fundamental Assumptions underlying distributed
  systems technology has changed
• wired gt wireless (limited range, high error
  rates)
• unlimited power gt minimize power
• Non-real-time gt real-time
• fixed set of resources gt resources being
  added/deleted
• each node important gt aggregate performance
• New solutions necessary

14
Example Resource Management

• Measure communication errors
• if too many
• increase communication power or if a mobile node
  it might move closer to the destination

15
Example Consensus

• Classical consensus all correct processes agree
  on one value
• No power constraints
• No real-time constraints
• Does not scale well to dense networks
• Approximate agreement (some work here) - on sets
  of values (physical quantities)
• New Solutions ?

16
New Concept of Consensus
Classical
New Definitions

• Termination every correct processor eventually
  decides some value
• Uniform Agreement no two processors decide
  differently
• Group Membership join/leave - everyone knows who
  is in the group

• Termination at least n correct processors
  decide some value by time t
• Group Agreement at least n processors decide the
  same value within epsilon
• Area/Function Membership join/leave an area or
  by function

17
Example Group Management (Tracking)
Base Station
18
Group Management - API

• Create_Group(name,function,criterion,atleast,accur
  acy) - implicit and explicit
• Destroy_Group(name)
• Join()
• Leave()
• Move_COG()
• Expand() -- to gain sensing confidence
• Shrink() -- to save power
• Commit(grp_ID) - to synchronize group
  re-configurations

19
Whats Hard

• Multiple targets
• Crossing targets
• False Alarms
• Depends on (changing) environment, sensors,
  confidence tradeoffs, noise, lost messages, )
• Speed of targets
• Uniqueness of targets
• Classify targets
• Proper abstractions
• Save power/min. commun.

20
The Essence

• Power
• Other limited resources (BW, CPU, )
• Extreme Scale
• Changing everything / uncertainty
• Aggregation
• unimportant individual nodes
• decentralized, very simple algorithms
• What I do impacts you (collisions) mutual
  exclusion

21
Six Themes

• Routing
• Power
• Localization
• Security
• Paradigms
• Theory
• Are there others? Yes..

22
Routing

• Solutions must be
• Power aware
• Robust to lost messages, dead motes, voids
• Real-time
• Communication range variations
• Moving end points
• Amount of state information
• Extreme Scale
• Secure

23
Power

• Example Algorithms
• AFECA power up and power down with time
  proportional to the number of neighbors
• GAF create grid and keep at least one mote
  alive in each grid (rotate among them in the
  grid)
• SBPM no grids non-deterministic minimize
  connectivity decentralized complete sensing
  coverage (60 savings over no power management)
• Differentiated Surveillance
• 50 less energy than best other solution

24
Power

• Other power savings
• Vary transmission power
• Turn off devices not needed
• On all devices on
• Off microprocessor in low power state so that
  registers/memory are not lost and clock interrupt
  can occur
• Checking microprocessor and radio are on
• Choose routes that minimize power
• Aggregate messages to save power

25
Localization

• Space (localization) and Time (clock sync) Basis
• Environmental monitoring where and when events
  occurred
• Localization is a function of
• Hardware available, cost requirement, signal
  propagation model, timing and energy
  requirements, network makeup, nature of
  environment, node and beacon density, time sync,
  communication costs, error requirements, device
  mobility,

26
Security

• What is the single most important issue that
  could prevent WSNs from wide scale deployment?
• Security
• 2nd issue -gt Privacy
• At application level
• Authenticity and integrity
• Security of each service (examples)
• Routing
• non-secure if a single node is captured!
• Eavesdrop or change message
• Flood
• Insidious unintended consequences of collecting
  data
• Monitor oceans for fish migration (data mine
  location of submarine fleet)

27
Security

• Localization
• Attacker can report he is close to everyone
• Chirp then wait then transmit to give false
  location (normally chirp and transmit
  simultaneously measure signals difference)
• Network Discovery
• Provide false messages to create false topology
• Prevent convergence

28
Paradigms

• Virtual Machines
• SQL and data services models
• EnviroTrack
• Tie to physical systems/physics
• Swarm computing
• Biological metaphors

29
Theory

• Theory of computation for WSN
• Emerging behavior of local/decentralized
  algorithms
• New graph theory
• New spatial-temporal analysis
• Aggregate control theory
• Utilization Equivalent Bounds
• Modeling and Analysis
• What are the fundamental scientific questions

30
Other Key Issues (1)

• Sensing/communication range ratio
• Sensing/communication/power tradeoffs

Communication Range
Sensing Range
What if the opposite?
31
Other Key Issues (2)

• Reality programming
• Robust to faults
• Sensor realities
• Dont believe one reading
• Hysteresis
• Sensor fusion
• Activation delays
• Avoid false alarms
• Self-Calibration

32
Other Key Issues (3)

• Limited capacities
• Rapid dynamics
• Scaling factors and implications on behaviors
• Extreme scaling
• Insidious interactions
• High density with many motes off to enable long
  system lifetime turn on when activity happens
  then too many with many collisions and poor
  response

33
Other Key Issues (4)

• Architecture hierarchy of control/capability/fun
  ctionality
• Size of targets/events (point/area)

X
Explosion
Fire
34
Middleware Services

• Non-traditional
• Configuration service
• Automatic calibration
• Network programming
• Reset services
• Management services

35
Middleware Services

• Real-Time Routing
• SPEED spatial-temporal concept
• Application Independent Data Aggregation
• AIDA feedback control
• Localization
• APIT realities of wireless world

36
Sensor Net Routing

• End-to-end
• Real-time
• Collisions
• Congestion

Destination
Source
Assumption Nodes know location
37
SPEED
USE VELOCITY
38
Application Independent Data Aggregation

• Expensive to acquire the channel
• Small data packets
• Group data packets into 1 MAC packet
• Works in addition to other data aggregation
  techniques which are based on semantics

39
Major Architectural Difference
40
FIXED SCHEME

• Accumulate N packets
• N degree of aggregation
• FIXED
• On Demand
• Adaptive/FC
• T Time out for old packets when accumulation
  rate is slow

41
DYNAMIC/Adaptive FC

• Adaptive choice of N
• Take into account the output Queue delay
• Delay is used to adjust the output queue push
  rate and degree of aggregation

42
Localization

• Determine the geographic location of each node
  with a high degree of accuracy (necessary for
  application)
• Applications
• search and rescue
• disaster relief
• target tracking
• Protocols
• location aware routing
• guaranteeing sensing coverage
• location directory services
• Fundamental and Enabling Service

43
Radio Model in Evaluation
DOI 0.2
DOI 0.05
Radio Model DOI Degree of Irregularity
44
Known Signal strength is not good indicator of
distance over the entire region Hypothesis
Signal strength IS accurate enough for nodes
very close to each other!
X
45
Testing Hypothesis
46
Summary

• (Much) Current Distributed Systems Technology
• wired networks, powerful nodes, highly reliable
  nodes, interaction with users, fixed numbers of
  resources/team members, unlimited power, ...
• Embedded (Large Scale) Distributed Systems
• wireless, simple nodes, unreliable nodes,
  interaction with the environment, resources being
  added and deleted continuously, power management
  needed,