SeRLoc: Secure Range-Independent Localization forWireless Sensor Networks

Presenter: Yawen Wei

Author: Loukas Lazos and Radha Poovendran Network Security Lab, Dept. of EE, University of Washington, Seattle, WA

Problem: Enabling nodes of a wireless sensor network to determine their location even in the presence of malicious adversaries.

Design Goal: Decentralized implementation Resource efficiency Robustness against security threats

Outline

Overview SeRloc Algorithm Comparison & Simulation Conclusion

Overview

What is Localization System? How to determine position? (Schemes)

1.range-dependent

GPS, Active Badge, Active Bat, Cricket

2.range-independent

DV-hop, amorphous localization, APIT, Centroid Why security?

Range-dependent Location System

Time of arrival (TOA) Angle of arrival Signal strength RF, acoustic, infrared and

ultrasound Disadvantages

(x1,y1); (x2,y2);(x3,y3)

SeRloc Algorithm

Network Model Location Determination Security Scheme Threat Analysis

Network Model N: unknown locations L: known locations and orientations, “locators” spatial homogeneous Poisson point process

SeRloc: Location Determination Secure Range-Independent Localization

a. the locator’s coordinates

b. the angles of the antenna boundary lines

c. R: the locator-sensor-communication range

d. Overlapping region

e. CoG (Center of Gravity)

Step 1: Locators heard

Step 2: Search area

Determination of the search area

A rectangular area of size less than

A rectangular area of size greater than

Step 3: Overlapping region-Majority vote

Step 4: Location estimation

SeRloc: Security Scheme Encryption:

Global symmetric key Sensor s & locator shares pairwise key

Locator ID authenticationCollision-resistant hash function (e.g. MD5)

Beacon of locator Li

SeRloc: Threat Analysis Types: Wormhole / Sybil / Compromised nodes Wormhole attack

Packet leshes: geographical / temporal Time measuring in challenge-reply scheme

Wormholes1. Sector uniqueness property

: area of locators heard by origin point

: area of locators heard by s

Wormholes2. Communication range violation property

A sensor cannot hear two locators that are more than 2R apart

the first message: q = k − 1 the last message: q = 0 All locators wait for a q Ts ∗ time

SeRloc: Threat Analysis Sybil Attack and Compromised nodes

Multiple network entitiesAssume sensor identitiesAssume locators (Not directly heard)

Outline

Overview SeRloc Algorithm Comparison & Simulation Conclusion

Comparison

Dv-hop and Amorphous localization

Comparison

APIT localization

5

Simulation Localization Error vs. Locators heard Localization Error vs. Antenna Sectors Localization Error vs. Sector Error Localization Error vs. GPS Error Communication Cost vs. Locators Heard

Conclusion

Secure localization in WSN Range independent Decentralized Security mechanisms Threats Higher accuracy and fewer locators