The Intrusion Detection in Mobile Sensor Network

1
The Intrusion Detection in Mobile Sensor Network
The Hong Kong University of Science and Technology

• Gabriel Y. Keung , Bo Li , Qian Zhang

Reporter ???
2011-5-18
2
Outline

• Introduction
• Network and Mobility Model
• The Intrusion Detection Problem in a MSN
• Sensitivity Analysis
• Simulation
• Conclusions Further Research

3
Outline

• Introduction
• Network and Mobility Model
• The Intrusion Detection Problem in a MSN
• Sensitivity Analysis
• Simulation
• Conclusions Further Research

4
Wireless Sensor Network(WSN)

• Consists of sensors for monitoring
• Temperature
• Vibration
• Pressure
• Motion
• Pollutants

• Cooperatively pass their data through the network
  to a main location.

• Each sensor is connected to one or several
  sensors
• a radio transceiver
• a microcontroller
• an electronic circuit for interfacing with the
  sensors
• an energy source.

5
Intrusion Detection in stationary WSN
Sensors and Cameras

• Detecting ground vibrations from foot steps and
  identifying warm body movement.

Mexican Side

• The operation of American Border Patrol
• places sensors along the American/Mexican borders
  (virtual fence")

US Side
6
k-Barrier Coverage(k-????)

• Barrier coverage
• treat sensor as barrier
• coverage be detected by a sensor

• Detect any intruders along any paths by at
  least k distinct sensors
• A very important concept for intrusion detection
  sensitivity analyze

7
Mobility Sensor Network(MSN)

• In stationary WSN
• sensors remain stationary after the initial
  deployment
• barrier coverage is determined by the initial
  network configuration and sensor deployment
• vacuum zone

• In MSN
• sensors can be relocated after deployment
• a better barrier coverage

8
The objectives of this paper

• characterize the k-barrier coverage
• compute probability of k-barrier coverage
• investigate the probability of k-barrier coverage
  sensitivity under different parameters
• examine the average travel distance

9
Outline

• Introduction
• Network and Mobility Model
• The Intrusion Detection Problem in a MSN
• Sensitivity Analysis
• Simulation
• Conclusions Further Research

10
The Kinetic Theory of Gas Molecules

• air molecule ? mobile sensor
• electron ? intruder
• The mean free path
• free pathAB,AC
• The average travel distance
• derived from kinetic theory(the mean free path)
• compute the probability of k-barrier coverage
  performance

B
A
C
11
Scenario

• A a long and narrow belt-like region
• N(A) the number of mobile sensors
• A the area of the region W the width of the
  area A/W the length
• nA the density of mobile sensors nAN(A)/A
• Rsensing range
• the sensor location can be modeled by a
  stationary two-dimensional Poisson process .

12
Mobility Model

• Sensors move independently
• Random direction mobility model
• Randomly chooses a direction ? ?0 2p ) with
  probability density function PT(?)
• Randomly chooses a speed from a range vm
  ?0,vmax, with probability density function
  PVm(vm)
• Once the boundary is reached, the sensor bounds
  back, by choosing another angular direction and
  continues the process.
• Intruder movement is crossing from one parallel
  boundary to another
• vi the velocity of an intruder

13
Coverage Measurement

• k-barrier coverage for an intruder traveling path
• k-barrier coverage for a mobile sensor network
• ? the cumulative coverage count by mobile
  sensors for any intruder paths
• Pr(?k) probability that a MSN satisfies this
  k-barrier coverage definition
• uncovered distance
• coverage rate (T\ Ts\ Tv) number of sensor
  coverage per unit time

14
Outline

• Introduction
• Network and Mobility Model
• The Intrusion Detection Problem in a MSN
• Sensitivity Analysis
• Simulation
• Conclusions Further Research

15
What we need?

• Goal Pr(?k)
• Kinetic theory of gas molecules
• average travel distance
• ? average uncovered distance
• Pr(?k) can be achieved by formulating the
  uncovered distance(?) and sensor coverage
  rate(Ts)

16
First assume sensors are stationary

• cross section of coverage

• the number of sensor coverage
• nAA Ts t
• average uncovered distance ?
• travel distance of an intruder divided by the
  number of sensor coverage

17
In mobile sensor network

• Relative speed vrel
• The coverage rate can be obtained by

18
The k-Barrier Coverage in an MSN

• Probability of k-barrier coverage in an MSN

19

• Average relative speed
• Expected total number of sensor coverage

20
Uncovered Distance Distribution

• the number of uncovered intruders t 0, N0 at
  time t, N
• during dt, the NTvdt intruders will be covered
• l uncovered distance for the length vit

21
Result Formulas
22
Outline

• Introduction
• Network and Mobility Model
• The Intrusion Detection Problem in a MSN
• Sensitivity Analysis
• Simulation
• Conclusions Further Research

23
Sensitivity analysis

• Density of Sensors and k-barrier Coverage
  Probability(nA , Pr(?k))
• Simplify the equations for analysis
• vi vm
• A50100, R1

24
Sensitivity analysis

• In a hybrid sensor network
• mobile static ?

25
Sensitivity analysis

• Sensor Speed and k-barrier Coverage Probability
• vm vi

26
Outline

• Introduction
• Network and Mobility Model
• The Intrusion Detection Problem in a MSN
• Sensitivity Analysis
• Simulation
• Conclusions Further Research

27
Simulation
28
Outline

• Introduction
• Network and Mobility Model
• The Intrusion Detection Problem in a MSN
• Sensitivity Analysis
• Simulation
• Conclusions Further Research

29
Conclusions Further Research

• Mobility can be exploited to obtain better
  barrier coverage
• Further Research
• detection error under varying sensor speeds
• study the optimal patrol route of controlled
  mobile sensors

30

• Thank You!
• Any questions?

Yu Zhitian
2011-5-18