Empiricalbased Analysis of a Cooperative LocationSensing System - PowerPoint PPT Presentation

Title: Empiricalbased Analysis of a Cooperative LocationSensing System


1
Empirical-based Analysis of a Cooperative
Location-Sensing System

Maria Papadopouli 1,2
1 Institute of Computer Science, Foundation for
Research Technology-Hellas (FORTH) 2 Department
of Computer Science, University of Crete
http//www.ics.forth.
gr/mobile/
This research was partially supported by EU with
a Marie Curie IRG and the Greek General
Secretariat for Research and Technology.
2
(No Transcript)
3
Overview

• Motivation
• Taxonomy of location-sensing systems
• Collaborative Location Sensing (CLS)
• Performance analysis
• Conclusions
• Future work

4
Motivation

• Emergence of location-based services in several
  areas
• transportation entertainment industries
• emergency situations
• assistive technology
• ? Location-sensing is critical for the support of
  location-based services

5
Taxonomy of location-sensing systems

• Modalities
• Dependence on use of specialized infrastructure
  hardware
• Position and coordination system description
• Cost, accuracy precision requirements
• Localized or remote computations
• Device identification, classification or
  recognition
• Models algorithms for estimating distances,
  orientation position

• Radio (Radar, Ubisense, Ekahau),
• Infrared (Active Badge)
• Ultrasonic (Cricket)
• Bluetooth
• Vision (EasyLiving)
• Physical contact with pressure (smart floor) or
  touch sensors

6
Cooperative Location-Sensing (CLS)

• Enables a device to determine its location in
  self-organizing manner using the peer-to-peer
  paradigm
• Employs a grid-based representation of the
  physical space
• ? can incorporate contextual information to
  improve its estimates
• Uses a probabilistic-based framework
• Each cell of the grid has a value that indicates
  likelihood that the local device is in that cell
• These values are computed iteratively using
  distance between peers and position predictions

7
Classifying CLS

• Modalities
• Dependence on use of specialized infrastructure
  hardware
• Position and coordination system description
• Cost, accuracy precision requirements
• Localized or remote computations
• Device identification, classification or
  recognition
• Models algorithms for estimating distances,
  orientation position

• Radio and/or Bluetooth
• Can be extended to incorporate other type of
  modalities

• No need for specialized hardware or
  infrastructure
• Can use only IEEE802.11 APs, if necessary

Grid representation of the space Transformation
to/from any coordination system Position a cell
in the grid
Objective 0.5 to 2.5 m (90)

• Computations can be performed remotely or at the
  device
• depending on the device capabilities

• Does not perform any of these functionalities

• Statistical analysis and particle filters on
  signal strength measurements collected from
  packets exchanged with other peers

8
Example of voting process (1/2)

• Accumulation of votes on grid cells of host at
  different time steps

9
Example of voting process (2/2)
Peers A, B, C have positioned themselves
Most likely position
Host A
x
x
x
Host C votes
Host B votes
10
Voting algorithm

• Initialize the values of the cells in the grid of
  the local device
• Gather position information from peers
• Record measurements from these received messages
• Transform this information to probability of
  being at a certain cell of its local grid
• Add this probability to the existing value that
  this cell had from previous steps
• Assess if the maximal value of the cells in the
  grid is sufficient high to indicate the position
  of the device

11
Example of training run-time signature
comparison
comparison
12
Position estimation (at peer A)
Landmarks vote

• Initialize the values of the cells in the grid of
  the local device
• Training phase Build a signal-strength map of
  the space (training-phase signatures)
• Run-time phase Build signal-strength signature
  of the current position
• Compare the run-time and training phase
  signatures
• For each new peer that sends its position
  estimation (e.g., peer B)
• Position B on the local grid of A based on Bs
  estimation
• Determine their distance based on signal-strength
  signature
• Infer likely positions of A
• Update the value of the cells accordingly
• Assess maximal weight of the cells, accept or
  reject the solution

Non-landmark peers vote
13
Signature based on confidence interval of
signal-strength values

• Weight of cell c assigned as

14
Example of confidence interval-based
comparison
T-, T confidence interval based on signal
strength measurements from an AP
T2

-
R
1

-

R
R2
R2
1
15
Distance estimation between two peers
16
Signature based on percentiles of the
signal-strength values
17
Particle filter-based framework

• step 1
• for L 1, , P
• (L-th particle)
• Transition
• Draw new sample xk(L) , P( xk(L)
  xk-1(L) )
• Compute weight wk(L) of xk(L), wk(L)
  wk-1(L) P( yk xk(L) ),
• where yk measurement vector signal
  strength values
• end loop
• Normalize weights
• Resample
• Goto step 1

18
Performance evaluation

• Performance analysis of CLS via simulations
  percom04
• Empirical-based measurements in different areas
• Various criteria for comparing the training phase
  and run-time signatures
• Particle-filter model
• Impact of the number of signal strength
  measurements
• Impact of the number of APs and peers
• CLS vs. Ekahau

19
Testbed description

• Area 7m x 12m _at_ Telecommunication and Networks
  Lab (in FORTH)
• Each cell of 50cm x 50cm
• Total 11 IEEE802.11 APs in the area
• 3.5 APs, on average _at_ any cell

20
CLS variations
21
Similarities between CLS Ekahau v3.1

• Use IEEE802.11 infrastructure
• Create map with callibration data
• Compare trainning run-time measurements

22
Ekahau vs. CLS

• no peers
• only APs participate

additional measurements
Percentiles capture more information about the
distribution of signal strength
23
Impact of number of APs
One AP off
24
Impact of peers
One extra peer
25
Use of Bluetooth instead of IEEE802.11
26
Conclusions

• The density of landmarks and peers has a dominant
  impact on positioning
• Experiments were repeated at the lab in FORTH and
  in a conference room _at_ ACM Mobicom
• median location error 1.8 m
• Incorporation of Bluetooth measurements to
  improve performance
• median location error 1.4 m

27
Discussion future work (1/2)

• Reduce training, management calibration
  overhead
• Easily detect changes of the environment
• density and movement of users or objects
• new/rogue APs
• Inaccurate information measurements
• Singular spectrum analysis of signal strength
• Distinguish the deterministic and noisy
  components
• Construct training and run-time signatures based
  on the deterministic part

28
Discussion future work (2/2)

• Incorporate heuristics
• about hotspot areas, user presence and mobility
  information, and topological information of the
  area (e.g., existence of walls)
• Experiment with other wireless technologies
• Sensors, cameras, and RF tags

I
29
UNC/FORTH Archive

• ? Online repository of models, tools, and traces
• Packet header, SNMP, SYSLOG, signal quality
• http//netserver.ics.forth.gr/datatraces/
• ? Free login/ password to access it
• Joint effort of Mobile Computing
  Groups _at_ FORTH UNC
• ? maria_at_csd.uoc.gr
• Thank You!
  Any questions?

30
Multimedia Travel Journal Tool

• Novel p2p location-based application for visitors
• Allow multimedia file sharing among mobile users

31
(No Transcript)
32
(No Transcript)
33
Simulations
34
Simulations

• Simulation setting (ns-2)
• 10 landmarks
• 90 stationary nodes
• avg connectivity degree 10
• transmission range (R) 20m
• For low connectivity degree or few landmarks
• the location error is bad
• For 10 or more landmarks and connectivity degree
  of at least 7
• the location error is reduced considerably

35
Bluetooth estimation experiments
36
Bluetooth-only estimationvalidation experiments
37
Joint IEEE802.11 Bluetooth estimation
experiments
38
Joint IEEE802.11 Bluetooth estimation
experimentsimpact of modalities - performance
analysis
39
Modality comparison