Localization in Wireless LANs

1
Localization in Wireless LANs
2
Outline

• Wireless LAN fundamentals
• Wi-Fi Scanner
• WLAN Localization
• Simple Point Matching
• Area Based Probability

3
Wireless LAN

• Standard
• IEEE 802.11a
• IEEE 802.11b
• Also call Wi-Fi
• operating at 2.4 GHz
• 11 Mbps
• IEEE 802.11g
• operating at 2.4 GHz
• 54 Mbps
• Future Standard

4
Wireless LAN

• Information
• MAC Address
• Identifier of the Wireless LAN Access Point (AP)
• Provided by the Ethernet LAN in the AP
• RSSI
• Signal Strength
• SSID
• Name of AP

5
Wi-Fi Scanner

• Platform
• Pocket PC 2003(Windows CE 4.0)
• Wi-Fi Network
• IEEE 802.11b
• API
• Windows CE .NET 4.2
• Tools
• Embedded Visual C 4.0
• Visual Studio .NET 2003

6
Wi-Fi Scanner

• Target
• Get two unique information
• MAC Address
• Signal Strength
• Future Application
• Develop the 2D Location Algorithm
• Provide the Multimedia Services (e.g. streaming
  service)

7
Wi-Fi Scanner

• Overall Architecture

User Interfcae
Core Operations
Embedded VC
Visual .NET 2003
DLL (dynamic link library )
8
Wi-Fi Scanner
Wi-Fi Application
Application
Presentation
Session
Transport
Network
Data Link
Physical
Network Driver Interface Specification (NDIS)
MAC Address, Signal Strength
9
NDIS

• Develop the Network Driver
• Support varieties of Network Technology (e.g.
  Ethernet (IEEE 802.3), Token Ring (IEEE 802.5),
  and IrDA media)
• Portability of drivers between platforms that
  support NDIS
• Network adapter miniport driver

10
NDIS Architecture
Miniport driver -communicate directly with
network interface card (NIC)
11
NDIS User-mode I/O(NDISUIO)

• Supports sending and receiving Ethernet frames
• Retrieve MediaSense indications
• Retrieve signal power indications

12
NDIS User-mode I/O

• Steps to bind the NIC Card
• Get Ethernet LAN Adapter (NIC) Name
• Create the File Handle to bind NDISUIO
• Using the DeviceIoControl interface to achieve
  the required packet

13
NDIS User-mode I/O

• fRetVal DeviceIoControl(
• hNdisUio,
• IOCTL_NDISUIO_QUERY_OID_VALUE,
• (LPVOID) pQueryOid,
• dwQueryBufferSize,
• (LPVOID) pQueryOid,
• dwQueryBufferSize,
• dwBytesReturned,
• NULL)

Retrieves NDIS object
14
NDIS User-mode I/O

• struct _NDIS_WLAN_BSSID
• ULONG Length
• NDIS_802_11_MAC_ADDRESS MacAddress
• Uchar Reserved2
• NDIS_802_11_SSID Ssid
• ULONG Privacy
• NDIS_802_11_RSSI Rssi
• NDIS_802_11_NETWORK_TYPE NetworkTypeInUse
• NDIS_802_11_CONFIGURATION Configuration
  NDIS_802_11_NETWORK_INFRASTRUCTURE
• InfrastructureMode
• NDIS_802_11_RATES SupportedRates
• NDIS_WLAN_BSSID, PNDIS_WLAN_BSSID

15
WLAN Localization

• Point-based approach
• Localization goal is to return a single point for
  the mobile object
• Area-based approach
• Localization goal is to return the possible
  locations of the mobile object as an area rather
  than a single point

16
Area-based Approach
17
Advantage of Area-based Approach

• Direct the user in the search for an object in a
  more systematic manner
• Presents the user an understanding of the system
  in a more natural and intuitive manner

18
Some Terms and Definitions

• n Access Points
• AP1, AP2, , APn
• Training set T0
• the offline measured signal strengths and
  locations an algorithm uses
• Consists of a set of fingerprints (Si) at m
  different areas Ai
• T0 ( Ai, Si ), i 1 m

19
Some Terms and Definitions

• Fingerprints Si
• Set of n signal strengths at Ai, one per each
  access point
• The are totally n access points
• Si (si1, , sin), where sij is the expected
  average signal strength from APj

20
Generating Training Set

• In one particular Ai, we read a series of signal
  strengths (sijk ) for a particular APj with a
  constant time between samples
• k 1 oij ,where oij is the number of samples
  from APj at Ai
• We estimate sij by averaging the series, sij1,
  sij2, sijo

21
Generating Training Set

• We do the same for all n APs, so we have the
  fingerprints at Ai,
• Si (si1, , sin)
• We do the same for all m areas, so we have the
  training set
• T0 ( Ai, Si ), i 1 m

22
Getting Testing Set

• The object to be localized collects a set of
  received signal strengths (RSS) when it is at
  certain location
• A testing set(St) is created similar to the
  fingerprints in the training set
• It is a set of average signal strengths from n
  APs, St (st1, , stn)

23
Area-based Approach Algorithms

• How to use the training set and testing set?
• Simple Point Matching
• Area Based Probability

24
Simple Point Matching

• Compare the received signal strength (RRS) in the
  training set and the testing set
• Find n set of areas that fall within a threshold
  of the RSS for each APj , j 1n
• The RSS with threshold for APj at position i
  sij q
• Return the areas formed by intersecting all
  matched areas from different AP area sets

25
Simple Point Matching

• How to choose the threshold?
• q is the standard deviation of signal received
  from the corresponding AP
• The algorithm starts with a very small q
• Area sets for some AP may be empty
• q is additively increases eg. q, 2q, 3q

26
SPM algorithm
27
Area Based Probability

• Goal is to return the area with the highest
  probability that the object is in
• Approach is to compute the likelihood of the
  testing set (St) that matches the fingerprint for
  each area (Si)

28
Area Based Probability

• Assumptions
• Signal received from different APs are
  independent
• For each APj, j 1n, the sequence of RSS sijk,
  k 1 oij, at each Ai in To is modeled as a
  Gaussian distribution

29
Bayes rule

• We compute the probability of being at different
  areas Ai, on given the testing set St
• P(Ai St) P(St Ai) P(Ai)/ P(St)
  (1)
• P(St) is a constant
• Assume the object is equally likely to be at any
  location. P(Ai) is a constant
• P(Ai St) cP(St Ai) (2)

30
Area Based Probability

• We compute P(St Ai) for every area Ai
  ,i1m,using the Gaussian assumption
• MaxP(Ai St) MaxcP(St Ai)
• MaxP(St Ai)
• Return the area Ai with top probability

31
Area Based Probability
32
Reference

• Eiman Elnahrawy, Xiaoyan Li, Richard P. Martin
  ,Using Area-based Presentations and Metrics for
  Localization Systems in Wireless LANs, Department
  of Computer Science, Rutgers University
• Andreas Haeberlen, Eliot Flannery, Andrew M.
  Ladd, Algis Rudys, Dan S. Wallach and Lydia E.
  Kavraki, Practical Robust Localization over
  Large-Scale 802.11 Wireless Networks, Rice
  University