Navigation Technology

 Wireless Localization and

Corresponding Techniques

Di He
Wireless signal properties, Classification of
common wireless localization methods

Basic principle of wireless localization method

based on angle and distance measurement

Basic principle of wireless localization method

based on fingerprint and proximity matching
 Contents
Wireless Signal Properties (Radio
Propagation)
Common Wireless Localization Methods
Common Wireless Localization Systems
Multiple-antenna Techniques
Cooperative Wireless Positioning
 Contents
Wireless Signal Properties (Radio
Propagation)
Common Wireless Localization Methods
Common Wireless Localization Systems
Multiple-antenna Techniques
Cooperative Wireless Positioning
 Wireless Signal Properties
Radio Propagation

Propagation Loss
 Wireless Signal Properties
The received wireless signal power varies
as a function of space, frequency and time.

The variation can be classified into either

large-scale fading or small-scale fading.
 Wireless Signal Properties

Path Loss: Propagation loss due to the distance between

transmitter and receiver
Large-scale Fading

Shadowing: Variations in the average signal strength due

to varying environmental clutter at different locations

Variations in time and space

Small-scale Fading

Frequency-selective fading
 Wireless Signal Properties
Main factors influence radio wave propagation:
(1)Reflection (反射): It happens when the
electromagnetic waves impinge upon a surface that has
dimensions much larger than their wavelength.
(2)Diffraction (衍射): Due to the effect of sharp edges in
the path of the radio waves between the transmitter and
the receiver.
(3)Scattering (散射): It happens when the
electromagnetic waves encounter objectives of
dimensions much smaller than their wavelength.
 Wireless Signal Properties
Large-scale Fading

Path Loss
Many models for predicting the path loss:
Okumura-Hata Model
COST231-Hata Model

while the above two models are fit for frequency up to

2GHz.
 Wireless Signal Properties
Okumura-Hata Model
L50 urban dB   69.55  26.16 log f c  13.82 log hre   hre   44.9  6.95 log hte log d
fc is the carrier frequency, hte is the height of transmission
antenna, d is the distance,  (hre) is the antenna height
correction factor.
For small and middle cities, we have
 h re   1.11log f c  0.7hre  1.5 log f c  0.8dB

For big cities, we have

 h re   8.29log 1.54hre 2 1.1dB, f c  300MHz
 h re   3.2log 11.75hre 2  4.97dB, f c  300MHz
 Wireless Signal Properties
So for the suburban area, the path loss
model becomes
L50 dB  L50 urban  2log  f c / 28  5.4
2

For the open terrain, the path loss model

becomes
L50 dB   L50 urban   4.78 log 2 f  18.33 log f  40.94
 Wireless Signal Properties
COST231-Hata Model
For the urban macro community environment,
the path loss model is
L  46.3  33.9 lg f  13.82 lg hte  44.9 lg d  6.55 lg hte lg d  1.11hte lg f
 0.7hte  1.56 lg f  3.8
For the suburban macro cell environment, the
path loss model is
L  46.3  33.9 lg f  13.82 lg hte  44.9 lg d  6.55lg hte lg d 
 3.2lg 11.75hte   4.97
2

while f is the transmission frequency with MHz.

 Wireless Signal Properties
And for the urban micro area environment,
the path loss is divided into two cases:

LOS:
PLurban_ macor _ LOS dB  32.4418  20 lg  f   20 lg d   0.0174d  20 lg max 0.013d / f ,1

NLOS:

PLurban_ macor _ LOS dB  min PLover _ the _ rooftop, PLBerg  
 Wireless Signal Properties
To overcome the problem the above two
models are fit for frequency up to 2GHz,
another model was proposed by Erceg et al.
For a close distance d0=100m, the median
path loss (ploss) in dB is given by

where ; shad is the shadowing

factor following a log-normal distribution with
a typical value around 6dB;  is the path
loss exponent given by
 Wireless Signal Properties
Values of a, b and c depend on the type of
terrain; hbs is the BS station height in
meter; and it is supposed that the receiver
antenna height is 2m, and the frequency is
2GHz.

For other frequencies and antenna height,

we have the modified path loss as
 Wireless Signal Properties
while pfreq is the frequency correction term:
(f: MHz)
pbsh is the receiver antenna height correction
term:
(2m<h<8m)

The propagation model used for LTE for

3GPP can be found in 3GPP protocol, where
different parameters are used for different
channel conditions and cell orientations.
 Wireless Signal Properties
Shadowing
Since the surrounding environmental clutter may
be different at different locations even with same
d, this variation is mainly due to reflection and
diffraction along the path of transmission, which
is called shadowing.
So the path loss can be expressed as

shad is a zero-mean Gaussian-distributed

random variable in dB, with standard
deviation shad.
 Wireless Signal Properties
Small-scale Fading
The receiving signal varies rapidly over a
short distance.

Reason:
(1)Multipath propagation
(2)Doppler frequency shift
 Wireless Signal Properties
(1) Multipath propagation
(θ1, β1, τ1)
X1(t)

z(t) g(t) XM(t)

path ith

Figure Multi-path transmission model of

wireless channel
 Wireless Signal Properties
The multipath parameters (amplitude ,
time delay and phase shift ) can be
characterized by a channel impulse response.
Suppose a unit impulse is transmitted and
there are nse scattering elements, we have

where is the Dirac delta function and j is

the imaginary unit of complex numbers.
 Wireless Signal Properties
So the channel impulse response can be
regarded as a function of time, frequency
and space.
When one of the multipath components
has a much higher power than the other
components, it is called Ricean channel;
When the above case doesn’t happen,
usually in NLOS scenarios, it is called
Rayleigh channel.
 Wireless Signal Properties
(2) Doppler frequency shift
The receiving signal frequency will be
different from the transmitting signal
frequency when there exists relative moving
between the transmitter and receiver.

receiving frequency:

where
 Wireless Signal Properties
Radio propagation and mobile
positioning

Measurements:

where is the 2D position of the mobile device;

is the ith known BS position;
is a function of ;
is the additive channel noise.
 Wireless Signal Properties

Typical measurements may include RSS,

TOA, TDOA, DOA (AOA) measurements,
database correlations (digital map
information), and position estimates.
 Wireless Signal Properties
Position Estimation

Given a measurement vector ,

the associated measurement transform

And assume the noise vector

where R is the autocorrelation matrix.

 Wireless Signal Properties
Then the position estimation may be obtained
by various algorithms:

(a)Nonlinear least-squares algorithm (NLIS)

When the noise is white, NLIS position
estimator can provide an optimal approach
equivalent to maximal-likelihood (ML) method
 Wireless Signal Properties

(b) Weighted NLIS

When  is a colored noise vector, NLIS can
be improved by the weighted NLIS estimate

(c) ML estimation
Generally when  obeys , the
ML estimate is formulated as

while the last term prevents large uncertainty.

 Wireless Signal Properties

Generally, when there is no closed-form

solution for above algorithms, we can use
the numerical search methods such as the
steepest-gradient method and Newton-
type methods. But they need a good
initialization, otherwise they may reach a
local optimum of the criterion function.
 Wireless Signal Properties

NLOS Positioning Error Mitigation

Above algorithms assume LOS signal exists,
while when NLOS signal also exists, it will
produce position error together with noise.
Generally, the pdf of NLOS measurement
error can be described as

where  is the mean bias, and is the

variance of the measurement error.
 Wireless Signal Properties

So the pdf of the measurement error can be

modeled by a mixture of two Gaussian
distributions as

where  stands for the probability of the

measurement resulting from LOS signal.
 Wireless Signal Properties

To mitigate the effects of NLOS path, two

methods are used:
(a)Use a robust error distribution (in the last
page) to design the positioning algorithm;
(b)Include some searching mechanism to
exclude the NLOS outliers.
 Contents
Wireless Signal Properties (Radio
Propagation)
Common Wireless Localization Methods
Common Wireless Localization Systems
Multiple-antenna Techniques
Cooperative Wireless Positioning
 Common Wireless
Localization Methods
Common Wireless Localization Methods

 Classification of Positioning Infrastructures

 Types of Measurements and Methods for their
Estimations
 Positioning Techniques
 Error Sources in Positioning
 Metrics of Location Accuracy
 Common Wireless
Localization Methods
Classification of Positioning Infrastructures
Positioning systems can be classified
according to different key characteristics.

Typical method concerns the system topology;

others may related to the properties of the
communication link; or related to the type of
integration into the communication
technology.
 Common Wireless
Localization Methods
(1) Positioning-system topology
 Common Wireless
Localization Methods
 Common Wireless
Localization Methods
(2) Physical coverage range

Problems:
Satellite----propagation in the ionosphere (电离层)
Cellular----NLOS, scattering, shadowing, multipath
Short-range----propagation statistics in different scenario
 Common Wireless
Localization Methods
(3) Integration of positioning solutions
 Common Wireless
Localization Methods
Common Wireless Localization Methods

 Classification of Positioning Infrastructures

 Types of Measurements and Methods for their
Estimations
 Positioning Techniques
 Error Sources in Positioning
 Metrics of Location Accuracy
 Types of Measurements and
Methods for their Estimation
Cell ID
It estimates the location of
the MSs as the position of
the fixed reference points
(FRPs) which the MSs are
connected to.

Disadvantage: In some
cellular networks the
range can be some few
Cell ID Technique thousands of meters.
 Types of Measurements and
Methods for their Estimation
Received Signal Strength (RSS)
The target MS measures the received signal
strength (RSS) from surrounding BSs.
Let denote the target MS position,
denote the positions of n
BSs, the RSS from the ith BS (BSi) is

Where di is the distance between BSi and

MS, pi is the received RSS at BSi.
 Types of Measurements and
Methods for their Estimation
In order to estimate the MS position, the
following least-squares objective function
can be defined as

Or in matrix form

where
 Types of Measurements and
Methods for their Estimation
Time of Arrival (TOA)
The TOA is the timestamp that a receiver
sees on its internal clock when a signal is
received at its terminals.
So the factor mostly influences the error is
the granularity (颗粒度) of the internal clock.
The maximal error:

c is the speed of light;

f is the clock frequency.
 Types of Measurements and
Methods for their Estimation
The TOA model is often described as:

where ttx is the timestamp in the transmitter clock

when the signal is transmitted; d/c is the real
transmission time; sync is a deterministic quantity
that compensates any mis-synchronization
between the transmitter and the receiver, that is:
 Types of Measurements and
Methods for their Estimation
Disadvantage: sync is often not possible to
obtain. Clock-synchronization in a whole
network is commonly complex to implement
and expensive to maintain.
Besides this, the propagation effects also
strongly depend on the scenario. So the
error could be generalized and simplified
into a Gaussian distribution as:
 Types of Measurements and
Methods for their Estimation
Time Difference of Arrival (TDOA)
The TDOA is the difference between two TOA
measurements obtained from two equivalent
signals emitted at exactly the same time.

d1-d2

d1
The TDOA t seen at
d2
the MS is given by
d3
d1-d3
 Types of Measurements and
Methods for their Estimation
Disadvantage: The BSs need to be clock
synchronized, though BSs and MSs don’t
need to be reciprocally (相互地) synchronized.
Introducing a variable component into the
TDOA is straightforward as:

where the two channel noise are supposed with equal

distributions, the extension of unequal distributions is
straightforward.
 Types of Measurements and
Methods for their Estimation
Direction of Arrival (Angle of Arrival)
A general model of the DOA (AOA) is to
assume an internal reference which all
angles and coordinates are related to.
Assuming the receiver is
placed at (x,y) and the
transmitter at (xtx,ytx), the
AOA  can be modeled as:
 Types of Measurements and
Methods for their Estimation
Regarding the noise component in AOA
measurement, it is generally assumed that it
has a zero-mean Gaussian distribution with
derivation aoa, the model of the AOA can be
written as:
 Types of Measurements and
Methods for their Estimation
Disadvantage:
 Types of Measurements and
Methods for their Estimation
Personal-information Indentification
 Airplane ticket
 Government census (人口普查)
data Data Mining
 Car identification information
 Credit card data
Another type  Data from security cameras Exact meaningful
of data  Computer-based information relationship to
(such as wireless MAC estimate or predict
addresses, login user names or movement trends
IP addresses) and patterns
 Personal-dependent
characteristics (such as
fingerprints, eye data or DNA
data)
 Common Wireless
Localization Methods
Common Wireless Localization Methods

 Classification of Positioning Infrastructures

 Types of Measurements and Methods for their
Estimations
 Positioning Techniques
 Error Sources in Positioning
 Metrics of Location Accuracy
 Common Wireless
Localization Methods
Proximity Sensing
(1)Physical contact

The MS can be localized when it establishes

physical contact with the BS. The BSs
may be some other sensors which is
possible to localize the user.
 Common Wireless
Localization Methods
(2) Identity methods
Assuming the MS detects n BSs
located at positions
the estimated MS position is:

When more information is available, the centroid algorithm can

be weighted either by a measurement of range or by the actual
communication range of each BS:
 Common Wireless
Localization Methods
(3) Macropositioning

Macropositioning concerns data-mining

techniques commonly used together with
personal-information identification, which are
not directly related to wireless positioning,
but represent an important group of methods
in the wide field of positioning.
 Common Wireless
Localization Methods
Triangulation (三角定位)
Lateration (三边定位法)
 Define the distance
between MS with
position and BSi
with position is
given by di, so:
 Common Wireless
Localization Methods
It is possible to obtain:

Or in the matrix form:

where

And the solution is given by:

 Common Wireless
Localization Methods
Hyperbolic localization (双曲线定位法)
Hyperbolic localization is an
alternative to the triangulation
technique using TDOA
measurements ti, resulting in a
difference of distances di
between BSi and BS1, that is:
 Common Wireless
Localization Methods
It can also be written as a matrix form as:

where

And the solution is given by:

 Common Wireless
Localization Methods
Angulation (测角定位法)
Assuming each BSi is
capable of measuring the
AOA i of the signal
transmitted from MS, it is
possible to obtain:
 Common Wireless
Localization Methods
And the matrix form is:

where

The solution is given by:

 Common Wireless
Localization Methods
Fingerprinting (指纹定位法)
 Common Wireless
Localization Methods
Centroid algorithm (质心算法)

In computational geometry, the

geometric center of a polygon (多
边形) is called the centroid (质心),
and the average value of the
polygon vertex (顶点) coordinates
is the coordinates of the centroid
nodes.
 Common Wireless
Localization Methods
Suppose the coordinates vector of the vertex
position of a polygon is expressed as p  ( x , y ) , then
i i i
T

the centroid coordinates of the polygon is as:

1 n 1 n 
( x , y )    xi ,  yi 
 n i 1 n i 1 

When the unknown node receives the number of

packet information from different beacon nodes
over a certain threshold, it can calculate the
centroid of polygons formed by these beacon
nodes, and determine its location.
 Common Wireless
Localization Methods
Advantage: Very easy to be realized with small
communication expenses.

Disadvantage: It is a kind of
localization method with coarse
precision. The calculation is based
on the assumption of spherical
wave (球面波) transmission. But
Figure. Real transmission
the real transmission power power contour of wireless
contour (等高线) may be different. signal
 Common Wireless
Localization Methods
DV-Hop Algorithm (distance vector-hop)
Mechanism: the unknown nodes and beacon
nodes first calculate the minimum number of
hops, and then estimate the average hop
distance; use the minimum number of hops
multiplied by the average hop distance to
estimate the distance between the unknown
nodes and beacon nodes, coordinate with the
lateration measurement or the maximum
likelihood estimation method to calculate the
unknown node.
 Common Wireless
Localization Methods
Calculation steps:
(1)Calculate the minimum hop number from
the unknown node to the beacon node
4 4
4
4
4
4 In the figure, A is the unknown
4 3 3 3 4 4
4
3
3 3
2
2 2
3
3
3
4 3
node, and the numbers represent
3 2 2
4
3
2
1 1
3
3
2
4 4
the minimum hop numbers from A
4 2 3

4
3
2 1
1
1
1
1 2
3
3
4
to certain beacon node.
3 4
2 1 A 3
2 4
4 3 1 1 4
4 1 1 2 3 3

4
3

3
2
2 1
1
1
2 3 4
4 It can be found the broadcast
2 2 2
3
3
2 2 2
2
2 3
3 4 packets are propagated in a
3
4
4
3 3
3
3
3 3 3
4
4
4
network by the same way as
4 4 4
concentric circles (同心圆).
Figure. Graphical representation of broadcast
packet propagation in beacon nodes
 Common Wireless
Localization Methods
(2) Calculate the actual hop distance between
the unknown node and a beacon node

Each beacon node estimates the real distance and

the average hop based on the location information
and distance hopping number recorded by other
beacon nodes in the first stage, using:

j i
( xi  x j ) 2  ( yi  y j ) 2
HopSize i 
h
j i
j
 Common Wireless
Localization Methods
where (xi,yi), (xj,yj) are the coordinates of the
beacon node i, j; hj is the hop number
between the beacon nodes i and j.

When the unknown node receives the

average per hop distance, the hop distance
of each beacon node is calculated according
to the number of hops of the record:
Di  hops  HopSize i
 Common Wireless
Localization Methods
(3) Self localization
The unknown node uses the lateration
method or maximum likelihood estimation to
calculate the position of its own.
 Common Wireless
Localization Methods
An example by using DV-Hop Algorithm:
(1)Suppose the real distances and hop
numbers within L1, L2 and L3 are
known. Then the average hop distance
for L2 is:
(40+75)/(2+5)=16.43 (m)
(2)Suppose the unknown node A get
the average hop distance information
from L2, then the distance from A to 3
beacon nodes are:
Figure. An example for DV-Hop L1&L3: 3x16.43(m), L2: 2x16.43(m)
(3)At last, the location of A could be
calculated by using the
lateration method.
 Common Wireless
Localization Methods
DV-Distance Algorithm
It is very similar to the DV-Hop algorithm.
The only difference is that it uses the RSSI to
calculate the distance between nodes, other
than the hop distance in above step 2.

Advantage: It can reduce the communication costs.

Disadvantage: It may introduce distance
measurement error by using RSSI.
 Common Wireless
Localization Methods
APIT Algorithm (approximate point-in-
triangulation test)
Mechanism: First, we determine a
number of triangular regions
containing unknown nodes. The
intersection (交集) of these triangle
regions is a polygon (多边形). It
determines a smaller area containing unknown
nodes, then calculates the centroid of the polygonal
region as the location of the unknown node.
 Common Wireless
Localization Methods
Theoretical Basis: Perfect point-in-triangulation
test (PIT).
A
A
M

M
B
C
C B

As shown in the above figure, If there is a direction,

when nodes move along this direction at the same
time away from or near the vertex A, B and C, then
node M is located outside triangle ABC; otherwise,
node M is located inside triangle ABC.
 Common Wireless
Localization Methods
Advantage: With the irregular propagation
mode of wireless signal and the random
deployment of sensor nodes, the positioning
accuracy of the APIT algorithm is high and
the performance is stable. And it has low
requirements for the distribution of beacon
nodes.
Disadvantage: APIT algorithm has high
requirements for network connectivity.
 Common Wireless
Localization Methods
Hybrid Approaches (混合定位法)
(1)Hybrid angulation and lateration

Simply merge the formulas of each methods

by Maximum Likelihood Estimation (MLE), it
can be defined as:
 Common Wireless
Localization Methods
where
 Common Wireless
Localization Methods
(2) Hybrid angulation and hyperbolic localization
It can be defined as:

where
 Common Wireless
Localization Methods
Common Wireless Localization Methods

 Classification of Positioning Infrastructures

 Types of Measurements and Methods for their
Estimations
 Positioning Techniques
 Error Sources in Positioning
 Metrics of Location Accuracy
 Common Wireless
Localization Methods
Error Sources in Positioning

Propagation
Geometry
Equipment and technology
 Common Wireless
Localization Methods
Propagation
It is the most common problems, and also
the most complex problems to deal with.
NLOS problem
Multipath fading problem
Shadowing
Body shadowing
Interference
The ionosphere (电离层)
 Common Wireless
Localization Methods
(1)NLOS problem
The NLOS paths problem is severe in
range-based systems.

For example, in the RSS-based method, the

attenuation will be:

where lobstr is the attenuation parameter

introduced by NLOS signal.
 Common Wireless
Localization Methods
It can be defined as a loss given by

Although the distance dobstr of propagation

through obstruction may be only a few
centimeters, the propagation parameter may
have a large propagation exponential obstr that
the attenuation can be of some tens of dB.
For this reason, it strongly influences the
positioning accuracy of range-based
systems in the indoor scenarios.
 Common Wireless
Localization Methods
(2) Multipath fading problem
With multipath fading, the multipath signal
received at the receiver from the same source
may have different attenuations, delays and
phases compared with the LOS signal.
Under NLOS conditions, it is common to model
the envelop of the received signal by a
Rayleigh distribution as:
 Common Wireless
Localization Methods
Given the relationship between the signal
and the received power is quadratic, the
Rayleigh distribution of the envelope results
in an exponential distribution of the received
power as:

So the model for the RSS can be written as:

 Common Wireless
Localization Methods
Under LOS conditions, the envelope of the
received signal follows a Ricean distribution:

where I0( ) is the modified Bessel function of

the first kind of order 0, and the parameter
a depends on the magnitude of the LOS.
 Common Wireless
Localization Methods
When a=0, the distribution reduces to a
Rayleigh distribution.
In this case the Ricean distribution of the
envelope results in a complex distribution of
the received power, which can be defined by
 Common Wireless
Localization Methods
(3) Shadowing
The measurements mostly affected by shadowing
are AOA measurements, although range-based
measurements are also affected. Then the AOA
model in static mode may be expressed by:

where bias is the bias introduced by the shadowing

factor.
 Common Wireless
Localization Methods
While in the moving case, the shadowing
effect shows a log-normal distribution as:

where  and  are the mean and standard

deviation of the variable .
And the expression in units of dB is given by
a normal distribution:
 Common Wireless
Localization Methods
(4) Body shadowing
Body shadowing is also a severe problem
for wireless positioning. It is due to the
present of the human body in close
proximity to the MS. It introduces additional
complex attenuations, and the impact of this
effect in wireless positioning is currently a
field of research under exploration.
 Common Wireless
Localization Methods
(5) Inteference
Inteference is another effect which has a
complex statistical treatment, and it strongly
depends on the systems active in the vicinity
of the MSs and the BSs.
Owing to the complexity of the problem, it is
commonly studied in very specific scenarios,
with strong assumptions, and it is typically
treated in terms of average values.
 Common Wireless
Localization Methods
(6) The ionosphere (电离层)
Ionospheric and tropospheric (对流层的)
effects are strongly present in satellite
positioning system.

The effects on the propagation also vary on

a periodic basis, assuming that additional
variations in solar activity are ignored.
 Common Wireless
Localization Methods
Geometry
Assume the position of the target to be X and
the position of the n BSs to be Xi
(i=1,2,…,n):

Suppose the clocks in BSs and MS have a

constant offset coff, which needs to be
calculated. Then the unknown position X is
extended with the clock offset coff:
 Common Wireless
Localization Methods
So it is possible to define the Euclidean
distance between each BS and MS as:

Suppose the measured ranges are

corrupted by a noise component :
 Common Wireless
Localization Methods
Expanding the term D(Y) as a Taylor series:

By defining
We have
Based on LS algorithm, the value for that
minimizes the error  is given by:
 Common Wireless
Localization Methods
Equipment and technology

Errors from equipment:

(1)Variability in the amplitude of errors
(2)Biases
(3)Non-stationary measurement processes
(4)Variations in the distribution of errors
(5)Thermal noise (most common type)
 Common Wireless
Localization Methods
Errors from technology:
(1)MAC protocols with randomized processes
(collision avoidance, listen-before-talk
techniques) can introduce additional delays
(2)Resolution of the measurements permitted
in the technology
(3)Clock synchronization errors in TOA and
TDOA
(4)Network management strategy
……
 Contents
Wireless Signal Properties (Radio
Propagation)
Common Wireless Localization Methods
Common Wireless Localization
Systems
Multiple-antenna Techniques
Cooperative Wireless Positioning
 Common Wireless
Localization Systems
Common Wireless Localization Systems
Infrared Positioning System WIFI Positioning System

Ultrasonic Positioning System Magnetic Field Positioning System

Bluetooth Positioning System Gyroscope Positioning System

RFID Positioning System Computer Vision Positioning System

…… ……
Indoor Positioning Schemes

Google Scheme

Nokia Scheme

IndoorAtlas Scheme

Duke University Scheme

 Google Scheme
2～3 satellites
signal could
be found even
indoor wifi
GPS
where there is
no GPS,
wifi or base
Blind station signals
Point

Google Maps
Base Floor Plan
Station Marker

……
 Google Scheme
Google has been working hard to solve two problems

Before After

Google is currently trying

to solve the problem of
Google tries to get
Get more data sources by means of
the plans and "crowdsourcing", which
Building information by is to encourage users to
plans itself. upload building plans.

When the users is After processing and

Collect some using Google analyzing in the server, a
GPS, wifi, indoor navigation, more accurate location
it will collect some service is provided for
BSs the user.
GPS, wifi, BSs
information information.
 Nokia Scheme
• When researching the indoor positioning technique,
Nokia Company developed a kind of device named “High
Accuracy Indoor Positioning (HAIP)”.

The device is equipped with a

ONE Bluetooth module and can be
positioned using this technique

A positioning launcher is installed in the

TWO room to complete the positioning through
the communication between the two.

The transmitter covers the range of 100m

*100m, with a precision of 30cm to 100cm,
THREE and the cost is low, and one or more
stations can realize the positioning.
 Nokia Scheme
3D positioning problem could be solved by utilizing multiple TXs,
and it can also improve the positioning precision and stability.
 IndoorAtlas Scheme

InDooRatlas is a mobile map The reinforced concrete structure

application developed by a of modern buildings will disturb the
research team at University geomagnetism in a local area, and
of Oulu of Finland. It can help the compass may also be
users navigate indoors by affected. In principle, a
identifying geomagnetism in heterogeneous magnetic field will
different locations. produce different magnetic field
observations because of its path.
IndoorAtlas locate by
identifying the unique Every building, every floor, even
geomagnetism of each every corner of the earth, their
position of the earth, and it geomagnetism is different. So,
can even work in a region they can be located by detecting
without wireless signals. their geomagnetism.

The research team has finished testing in a 4593 foot deep

(1380 meter) mine, which runs well. The precise range of the
application in the building is 0.1 meters -2 meters. Users can
use other supporting software and hardware, and can get more
precise location information with other map applications.
Duke University Indoor
Positioning Scheme
 Duke University Indoor
Positioning Scheme
Duke University uses the idea of landmark in
real life, such as WiFi, 3G signal dead angle,
ONE and some movement characteristics, such as
elevators, staircases, etc., and calculates the
location according to the known road signs.

When you move, follow the other sensors

UnLoc TWO (gyroscopes, accelerometers, azimuth
sensors, etc.) to track your position.

The accuracy of this process will be

THREE gradually reduced, but when you arrive at
the next sign, the position will be calibrated.
Main Problems in Wireless
Positioning Techniques
Now a lot of technology has been applied to wireless
positioning, such as GPS, RFID, Wi-Fi and so on.
However, they must use their own application interface
through their respective protocols. The development of the
heterogeneous scheme is seriously blocked.

High-precision requirement is still a main problem which

restricts the popularized and quickly application.

Security of location sharing is a serious problem.

 Application Prospect of
Indoor Positioning
With the popularity of smartphones and the development of
mobile Internet, maps and navigation software will enter a new
era - indoor navigation. Text Text
Text

The average time of people outdoor

in 80-90% is indoors

70% of the mobile phone

is used indoors
80% of the data connects indoor
are used indoors

2004
 Application Prospect of
Indoor Positioning

First, public security and emergency response. In an emergency,

everyone wants to be accurately positioned by the rescuers to the
location of the building, even the floor or room number.
Secondly, it can be applied to mobile shopping, mobile e-
commerce, personalized advertising / preferential information.
 Application Prospect of
Indoor Positioning
For large buildings, public safety, emergency evacuation and rescue after the
disaster of major applications, breakthroughs in key technologies in wireless
sensor network of large and complex buildings and precise indoor positioning
and location services based on large building complex environment indoor
positioning system is developed, and is widely used in the National Grand
Theater, to achieve seamless indoor and outdoor positioning precision and
provide technical support for large building navigation monitoring and
emergency service management.
 Contents
Wireless Signal Properties (Radio
Propagation)
Common Wireless Localization Methods
Common Wireless Localization Systems
Multiple-antenna Techniques
Cooperative Wireless Positioning
 Multiple-antenna Techniques
The use of multiple antennas at the
transmitter and/or receiver provides various
benefits, such as:
Array gain
Interference reduction
Diversity gain
Multiplexing gain
Simplify the multichannel equalization
particularly simple by using OFDM
 Multiple-antenna Techniques
Multi-antenna techniques can be broadly
classified into the following 3 categories:
(1)Smart array processing, such as
beamforming (BF), to increase received
power and reject unwanted interference;
(2)Spatial diversity (SD), to mitigate fading
and enhance link reliability;
(3)Spatial multiplexing, to boost the data
rate.
 Multiple-antenna Techniques
Spatial diversity
Two or more antennas physically seperated
from each other are used to obtain
independent versions of the received
signals. They can be located at the receiver
(receive diversity), at the transmitter
(transmit diversity) or at both sides.
Ideally, an antenna separation of half of the
wavelength (/2) can guarantee the spatial
fadings are uncorrelated.
 Multiple-antenna Techniques

Spatial diversity is relatively simple to implement

and does not require additional frequency spectrum.
When multiple copies of the transmitted signal arrive
at the receiver, they have to be precombined in
order to obtain a diversity gain.
 Multiple-antenna Techniques
Spatial multiplexing
 Multiple-antenna Techniques
At the qth receive antenna, the independent
source signals are linearly
combined as:

where hq,p is the channel fading between the

pth transmit and the qth receive antenna,
is the channel noise at the qth receive
antenna.
It can be certified that the source signals
can be estimated if nrx>ntx.
 Multiple-antenna Techniques
As a result, spatial multiplexing promises
higher throughput compared with single-
input-single-output (SISO) systems, without
the need to increase the system bandwidth
or the transmit power.

Under certain circumstances the system

capacity increases linearly with the number
of antenna used.
 Multiple-antenna Techniques
MIMO and mobile positioning

In some circumstances with

complex environment (such
as urban area), MIMO
technique can use information from multipath
components to improve the positioning
accuracy. And with the proper channel model,
the location problem can be solved using
sequential Monte Carlo methods, also known as
particle filtering.
 Contents
Wireless Signal Properties (Radio
Propagation)
Common Wireless Localization Methods
Common Wireless Localization Systems
Multiple-antenna Techniques
Cooperative Wireless Positioning
 Cooperative Wireless
Positioning
In the cooperative wireless communications,
the users may increase their effective QoS
via cooperation. In this case, each user is
assumed to transmit their own data and act
as a cooperative agent for other users.
trade-offs in
Code Rates
Cooperation
Transmit Power
 Cooperative Wireless
Positioning
Transmit both data Trade-off
Need more
from themselves
power
and others Hope to achieve a
net reduction of
the transmit power
Need less Baseline transmit
power power for all users
will be reduced

Transmit both bits Trade-off

Need more
from themselves
transmit rate
and others Hope to achieve a
net reduction of
the transmit rate
Need less Baseline transmit
transmit rate bits for all users
will be reduced
 Cooperative Wireless
Positioning
RSS-based cooperative positioning

Suppose MS1 is the target MS,

are the coordinates of the MSs,

are the coordinates of the BSs.

 Cooperative Wireless
Positioning
The following least-squares (LS) objective
function can be proposed:

The optimum result of J(X) gives a position

estimates for all MSs.
 Cooperative Wireless
Positioning
Cognitive Radio and Mobile Positioning

In the CR frameworks, there are two types

of users:
Primary users, who have priority to access
the spectrum;
Secondary users, who have no priority or
a lower priority than primary users.
 Cooperative Wireless
Positioning
Kim et al. proposed a scheme to be used by the
secondary users to estimate the unknown primary
users’ positions and transmission power, under the
following assumptions:
(1) The secondary users’ positions are known;
(2) The secondary users measure RSS values
from primary users;
(3) At least four secondary users receive a signal
from each primary user;
(4) The shadowing effect is independent for each
secondary user.
Cooperative Wireless
Positioning
 Cooperative Wireless
Positioning
Let the real position of the primary user be
, and the coordinates of the ith
secondary user be
Let di denote the distance between the
primary user and the ith secondary user:

The ideal RSS at the ith secondary user is:

 Cooperative Wireless
Positioning

Generally, we can refine the sample data by using the sample average:

where is the kth sample RSS value at the ith secondary user in dBm,
and m is a predefined number of samples.
 Cooperative Wireless
Positioning
So we can use a measurement model as:

This measurement model is based on the

sample mean of the RSS values at each node.
And it can reduce the shadowing disturbance in
the RSS values.
So the relationship between the measurement
and the propagation model is as:
 Cooperative Wireless
Positioning
If there is no shadowing effect in the RSS
measurement, we have

And we can obtain that

where
 Cooperative Wireless
Positioning
The above equation holds for each
secondary user that is receiving the primary
user’s signal. So it can also be written in
matrix form as:

where
 Cooperative Wireless
Positioning
The above estimation problem can be
solved by using the LS method, with the aim
of minimizing the disturbance caused by
shadowing to the measurements.
Let be the estimated position, the solution
can be computed as
 Summary
Basics of wireless communications and
how those are related to mobile positioning.
Classification methods for wireless
positioning.
Basic wireless positioning methods.
New trends based on multiple-antenna
technique and cooperative wireless
positioning.
Thank You ！