A Software Based Indoor Relative Location Management System

1
A Software Based Indoor Relative Location
Management System

• Zhe Guang Zhou, Aruna Seneviratne, Richard Chan
  and Prawit Chumchu
• In Proceeding of Wireless and Optical
  Communication, Canada 2002
• Present by Shwu-Fen Kuo

2
Outline

• Introduction
• Related work
• Relative location management
• Location estimation model
• Experimental evaluation
• Conclusions and future work

3
Introduction(1/2)

• Detecting and tracking the physical location of
  mobile devices
• Hardware scheme additional hardware to provide
  higher accuracy
• User need purchase additional hardware
• Cost increase and system configuration
  requirement
• Not yet standardized, hence usage is very limited
• Software scheme does not need hardware but
  acceptable accuracy
• Overcome these obstacles
• With standard WLAN equipments

4
Introduction(2/2)

• This paper, establish the limitations of software
  based location management, and a scheme that
  provides better accuracy
• Based on RF signal strength
• Use fixed devices in the operating environment
• With basic triangulation techniques

5
Related work(1/2)

• The hardware based systems
• Active badge system
• Using Infrared (IR) signal for position inference
• Hardware unitbadge
• Require high-density sensors
• Bat system
• Using RF and ultrasound signals
• Hardware unitBat (wireless transmitter)
• Cricket compass system
• Using RF and ultrasound signals
• Hardware unitCricket Compass (CC)

6
Related work(2/2)

• Hardware based location management system
• Provide accurate location information
• Scalabilities of the IR based system are poor
  because of the limited range of IR devices
• Require additional device for signal transmission
• Incurs the cost of installation and maintenance
• The software based system
• RADAR system
• The construction of the radio map
• The accuracy of radio map

7
Relative location management(1/3) Operational
environment

• The computer related equipment used in indoor
  area
• stationaryfixed locations (printers, PCs and
  Access Points)
• Mobile devicescarried around by individual to
  best suit the working environment
• The location of Mobile Hosts (MN) can be detected
  by using triangulation

8
Relative location management--(2/3) Combined
configuration

• The objective of the combined scheme
• Using infrastructure configuration for providing
  the MH to access network
• The Ad-hoc configuration is for detecting its
  location
• Figure 1
• Advantage
• Not necessary to consider the layout and
  partition in the office environment
• The system is independent of the type of building

9
Relative location management(3/3) Prototype
implementation

• The system operates only in the ad-hoc mode, and
  no data transfer between the MHs
• The prototype used four laptops
• Three emulation stationary devicesFixed Points
  (FPs)
• One as the MH
• Using the SS rather than SNR
• 4 samples/sec

10
Location estimation model (1/9) Indoor signal
propagation model

• Affect the signal propagation
• Multipath
• The layout of the building
• The construction material used
• Human movement
• The indoor signal path loss
• The indoor signal propagation model

11
Location estimation model (2/9) Line-of-Sight
(LOS) experiments

• Setting up two laptops in peer-to-peer mode
• One laptop acted as a FP to located at one end of
  the corridor
• The other acted as a MH moving away
• The SS varies with distance and also the
  variation due the orientation of the receive
• Figure2

12
Location estimation model (3/9) Line-of-Sight
(LOS) experiments

• Different orientation2.3dB12.5dB
• Human body6.4dB
• 15cm thick wall2.1dB
• ,AF is the attenuation factor
• An AF3 was used as it gives an acceptable
  approximation for attenuation due to walls
• Attenuation due to human (i.e. 6dBmAF,with m2)
• Attenuation due to the orientations ( 9dB in
  average with m3)

13
Location estimation model (4/9) Location
estimation

• The distance between the MH and FP is computed by
  the signal propagation equation
• The log file of the NIC driver provide
• Average SS is used to compute the distance from
  MH to FP
• Minimum and maximum SS indicate that there exists
  an upper and lower bound distance for each sample
  reading

14
Location estimation model (5/9) Location
estimation

• Ideal case
• This SS fits into the negative logarithm curve of
  the indoor propagation model
• Three distance circles from different FPs should
  intersect at one point the physical coordinates
  of the MH
• Angle ?, using the cosine rule since the distance
  from the MH to the FP0 and FP1
• The distance from MH to FP2 can be used to
  determine whether the MH is outside the triangle
  or not
• Figure 4

15
Location estimation model (6/9) Location
estimation

• Non-ideal case1
• The measured SS is smaller than the value on the
  curve with the same distance
• The estimated distance is slightly larger than
  the exact distance
• Using the centre of mass of the overlapping
  triangle as the location point of the MH
• Figure 3

16
Location estimation model (7/9) Location
estimation

• Non-ideal case2
• Scaling up the radius of each distance circle
• The centre of mass at the intersection of the
  triangle is taken as the location of the MH
• Figure 3

17
Location estimation model (8/9) Location
estimation

• Non-ideal case3
• Often happens in indoor, the actual attenuation
  is larger than the estimated attenuation
• The measured SS will be lower than what we
  expected
• Using the scaling factor m , as a multiple of
  attenuation factors (AF) to dynamically increase
  the path loss factor Xs scaling down the radius
  of the trangle
• Figure 3

18
Location estimation model (9/9) Location
estimation

• Non-ideal case4
• When the received SS fluctuates, even when the MH
  is stationary
• This SS fluctuation is due to the multipath
  phenomenon
• Circular buffer is used to store the recorded SS,
  the median value of the last ten SS samples are
  then used compute the current location
• Cause some delay in estimating the current
  location

19
Experimental evaluation(1/2)

• The actual distance from the MH to the FPs is
  compared with the estimated distance to determine
  the absolute error distance
• Twenty measurement point

20
Experimental evaluation(1/2)
21
Conclusions and future work

• This paper presented a software based relative
  location management scheme
• It will possible to provide an accuracy of less
  than 3m
• Without the burden of having to construct the
  Radio Maps
• Improving the robustness and accuracy
• Using the movement histories and exploiting the
  fact that the mobile devices can only move in a
  limited area

22
Combined configurations
Return
23
Line-of-Sight (LOS) experiments
Return
24
Location estimation Ideal case
Return
25
Location estimation Non-ideal case
Return