Locus: WLAN Location Sensing MQP

1
LocusWLAN Location Sensing MQP

• Arvinder Singh
• Ali Taheri
• Advisor Prof. Emmanuel O. Agu

2
From Star Trek to Locus
Star Trek The Next Generation (Paramount
Pictures 1987)
3
Project Goals

• Our Ideal MQP
• An original project involving RD
• Potential for product development and publication
• Objectives
• Determine physical location of clients on an
  802.11b wireless network
• High Quality Interactive Graphics
• Software-only, platform independent solution

4
Possible Applications

• Administrative monitoring tool
• Track loaned laptops and PDAs
• Campus Center, Library, and ATC
• Context aware applications
• Instant Messaging
• Library catalog
• Assisted tours
• Security
• Theft and intruder detection

5
Related Technologies

• Global Positioning System (GPS)
• Ekahau and Newbury Networks
• Wireless Research API (WRAPI)
• Windows based API
• Hardware independent access to Signal Strength
• Scalable Vector Graphics (SVG)
• Open standard similar to Macromedia Flash
• Ideal for displaying CAD files
• Zoom without distortion
• Platform independent Java viewer (Batik)

6
Location Sensing Taxonomy
7
Locus Subsystems

• Modular Design to allow
• WhereAmI (Single user mode)
• WhereAreYou (Multiple user mode)
• WhereAreThey (Administrative mode)

8
Step 1 Setup

• Configuration
• Acquire floor maps in SVG format
• Setup naming convention
• Enable AP broadcast messages
• Calibration
• Record signal strength tuples
• Associate tuples with points on map
• Record point-tuple association in database

9
Step 2 Calculate Location

• Algorithms
• Min. Threshold
• Euclidean Dist.
• Joint Probability
• Bayesian Filters

10
Step 3 Display Location

• Java UI
• SVG Layers
• Batik API

11
Results Locus 1.0

• Completed Modular OO Design
• Integrated WRAPI for retrieval of SS from
  hardware
• Analyzed platform independence issues of WRAPI
  and produced recommendations
• Integrated Batik for display of maps and location
  using SVG
• Implemented matching algorithm
• Successfully integrated modules allowing us to
  determine user location

12
Follow-up Work Locus 1.5

• Conference submissions
• SVG Open, Tokyo, Japan
• ADHOC-NOW, Vancouver, BC
• Independent Study Projects
• Compare matching algorithms
• Develop context aware applications
• Platform testing and quality assurance
• Industry Collaboration
• Intel Research collaboration possibilities being
  discussed

13
Future Work Locus 2.0

• Complete tweaks for release version
• Develop Network Subsystem
• Develop Linux and Mac agents
• Complete context-aware applications
• Implement WhereAreYou and WhereAreThey
  functionality

14
Acknowledgments

• Computer Science Department
• Prof. Emmanuel O. Agu
• Kutty S. Banerjee
• WPI Network Operations
• Sean M. O'Connor
• Joseph M. Krzeszewski
• WPI Plant Services
• John E. Miller
• UCSD Wireless Research API (WRAPI)
• Anand Balachandran

15
Questions