Location Awareness in Pervasive

Computing
School of Computer Science and
Engineering

Vellore Institute of Technology, Vellore

General Architecture: Location Aware
Computing
COO: Cell of Origin, AOA: Angle of Arrival, TOA: Time of Arrival, GPS:
Global Positioning System, LPM: Location Pattern Matching [Ref: Kim, Young
Jae. ”Location Aware Computing.” Computer Science and Engineering Department University of Texas
at Arlington.]
Mobile Phone Tracking

“Mobile phone tracking is a process for identifying the

location of a mobile phone, whether stationary or
moving. Mobile positioning may be used for location-
based services that disclose the actual coordinates of a
mobile phone. Telecommunication companies use this to
approximate the location of a mobile phone, and
thereby also its user.”

1
htt ps://en.wikipedia.org /wiki/Mobile_phone_tracking
Cell Of Origin (COO)

“Cell of origin (COO) is a mobile-positioning technique

for finding a caller’s cell (the basic geographical
coverage unit of a cellular telephone system) location. ”
Contd..

“Crude COO positioning considers the location of the

base station to be the location of the caller. This is not
very accurate, as the majority of mobile network cells are
projected from an antenna with a spread of 120◦ (i.e.
three mounted on a mast to give complete coverage)
giving a signal coverage area with the base station at
one corner, rather than the centre. Omnidirectional cells
may be used in rural locations (which typically have
large ranges and hence uncertain locations for phones
within them) and in cities (where they may have ranges
of a few hundred metres). ”
Time Of Arrival (TOA)

“The signal arrival time differences from the mobile to

more than one base station are used to calculate the
location of the device. This method needs
synchronization of cellular network using very expensive
atomic clocks at each cell sites. The cell sites use
Location Measurement Units (LMUs). By measuring the
signal from the mobile device, the LMUs can triangulate
the user’s position.”
Contd..

“The wireless device’s signal is received at various base

stations. Since each base station is at a different
distance from the device, the signal arrives at a slightly
different time. The technique requires signal timing
information from at least three different base stations.
The receivers, synchronized by the atomic clock, send the
user’s signal and timing data on to the LMU, where the
times are compared and computed to generate a latitude
and longitude for the user’s device by triangulation.”
Contd..

“The cost benefit analysis is not in favor of the usage of

this technology, since the cost of implementing this is
higher when compared with the enhancement in the
performance. While TOA is more accurate than COO
technology, it is expensive because of the large number
of LMUs required.”
Angle Of Arrival (AOA)

“AOA requires a complex antenna array at each cell site.

These antennas together are doing a major chunk of the
work to determine the angle from which a cellular signal
has originated. The wireless device’s signal is received
at the antennas of various base stations. Each site is
equipped with additional gear to detect the angle from
which the user’s signal has come. Generally, at least
three sites must receive the signal to provide an
accurate location. The receivers, then, send the
compass data to the LMU, where the angles are
compared and computed to generate a latitude and
longitude for the user. Since the only difference of AOA
from TOA is to use angles instead of time of flight, AOA
does not differ much from the TOA.”
TOA vs
AOA
Enhanced Observed Time Difference (E-OTD)

► “Enhanced Observed Time Difference (E-OTD) is a

standard for the location of mobile telephones. The
location method works by multilateration1.
► It involves time difference measurements being
made in the handset rather than the network, and
a mechanism to pseudo-synchronise the network.
► Enhanced observed time difference (E-OTD)
measures the time differences of arrival of a signal
from multiple(>=3) base stations, and can be
implemented without changing the network.”

1
1
1
Location Pattern Matching (LPM)

“LPM exploits the distortion of the signal caused by the

obstacles. The wireless device’s signal is reached at the
base station equipped with special gear, and the
receivers send the signal to the special Location
measurement unit(LMU) where sophisticated equipment
analyzes the acoustic radio signal and compares it to a
database of previously identified locations with standard
signal characteristics and a match is made. These
characteristics include signal reflections like multi-path,
echoes, and other signal anomalies. The LPM eliminates
the need for line of sight triangulation involving multiple
cell sites. This is particularly important in dense urban
environments where buildings tend to obstruct the line of
sight, and in sparse rural environments where it is
unlikely that three or more base stations will be available
to receive the signal.”
Global Position System (GPS)

“GPS is one of the global navigation satellite systems

(GNSS) that provides geolocation and time information to
a GPS receiver anywhere on or near the Earth where
there is an unobstructed line of sight to four or more GPS
satellites. It does not require the user to transmit any
data, and operates independently of any telephonic or
Internet reception, though these technologies can
enhance the usefulness of the GPS positioning
information.”
Contd..

“GPS is a passive one-way system where all signals are

transmitted by earth-orbiting satellites, and position
determination happens at the receivers”

1
htt ps://www.uio.no/studier/emner/matnat/ifi /IN5600/v23/
s l i d e s -lec tures/
Contd..
Advantages of
GPS

► “Worldwide coverage
► Scales to an unlimited number of users
► Preserves user privacy
► Supports a range of location services with accuracy
that ranges from several meters to a few millimeters
[low to high].”
Contd..
Disadvantages of
GPS

► “Accurate GPS localization requires an

unobstructed view of at least four satellites
► GPS signals do not penetrate well through walls,
soil, and water: Thus, the system cannot be used
inside buildings, underground (e.g., inside a mine or
tunnel), or for subsurface marine navigation
► Signal can also be obstructed by large
buildings in the so-called urban canyons
► Normally, location GPS based services require a
monthly extra-fee.”
GPS Architecture

“ It consists of three distinct parts:

► A constellation of Earth-orbiting satellites that
broadcast a continuous ranging signal
► Ground stations that update the coordinate
projections and clocks of the satellites
► the receivers that use the GPS signals to estimate
their position
Contd..

Satellites orbiting the Earth:

► 31 satellites organized
into six circular orbits
26,560 km above the
Earth with a 12-h
period.
► Full GPS coverage
requires 24 GPS
satellites
a
► The additional
satellites operate as Image taken from
a

Google
active spares to
accommodate
occasional
maintenance downtime
and to assure system
Contd
..

Ground Stations:
► Responsible for monitoring satellite positions and
providing satellites with clock corrections and
satellite orbit updates
► There are enough ground monitoring stations to
allow each satellite to be simultaneously tracked
by at least two monitoring stations
► Simultaneous satellite tracking improves the
precision of orbit calculations increasing localization
accuracy
Contd
..

GPS Receivers:
► Determine their position by simultaneously tracking
at least 4 and commonly up to 12 satellites
► Can be augmented with other sensors (e.g.,
altimeters, accelerometers, and gyroscopes) to
compensate for gaps in GPS coverage
Contd
..

Receivers:
► {In earlier times, receivers were bulky.} Today, they
are typically around the size of a cell phone, and
newer single chip GPS implementations have made
possible form factors as small as a wristwatch”
GPS
Algorithms

“Supported by all GPS receivers and by tracking four

or more satellites, allows the receiver to:
► Estimate its position in three dimensions (latitude,
longitude, altitude)
Contd..
How is the location estimated?

► Computed from the estimated positions of the

satellites
► The ranges from the receiver to those satellites
Contd
..
S1

► Satellite locations are

S2
(x_i, y_i, z_i)

R_i

learned from the

broadcasts from the
b
S3

(x,y,z)

satellites
► On the figure in the S4

right Ri is the distance

between the receiver
and the satellite S2
► Location is inferred by
measuring the transit
time of the signal
between the satellite
and the receiver and
multiplying it by the
speed of light in
Contd ..
Accurately Measuring the Signal Transit Time

► Measuring the signal transit time accurately

requires the satellite and the receiver’s clocks to
be tightly synchronized
Contd ..
Accurately Measuring the Signal Transit Time

In Practice:
► The use of various low-cost crystal oscillators in the
receiver introduces a bias that makes the distance
from the satellite appear shorter or longer than the
real value
► The receiver-induced clock bias will be the same
across all of the satellites
► The satellite induced clock bias is less of an issue
because of their extremely accurate atomic clocks
► Thus, the effect of the receiver’s clock bias can be
removed by treating it as an extra unknown in the
calculation of the location.
Contd..
Equation!

► The receiver’s location in the three-dimensional

space (x, y, z) and the receiver clock bias b is
determined by solving the following equation for at
least four satellites:

q
Ri = ( x i − x )2 + (yi − y )2 + (zi − z )2 − b

Why four satellites?

► In cases where more than four satellites signals are

available to the receiver, the redundant data is used
to try and identify error and eliminate it in the
location estimate. This is typically done using a
least-squares estimation or with a Kalman filter
Contd..
Satellite Range Estimation

► To allow the distance between the satellite and the

receiver to be estimated, GPS satellites transmit
radio signals modulated by Pseudo-Random Noise
(PRN) codes. Such codes consist of binary sequences
that appears to be randomly generated
Contd..
Satellite Range Estimation

► GPS receivers continually compare the signals they

are receiving with a locally generated replica of the
satellite’s PRN code. The time delta between the
received signal and the local PRN code represents
the signal’s travel time. Range is computed by
simply multiplying the travel time by the speed of
light in vacuum (299,729,458 m/s)
Contd..
Errors and
Biases

► Ionospheric delay is the dominant source of GPS ranging

error. It is caused by the interaction of the GPS signal
with ionized gases in the upper atmosphere. It varies
with time of day, time of year, solar flare activity,
and the angle of entrance of signal that affects the
length of the path through the ionosphere
Contd..
Errors and
Biases

► Satellite coordinates errors are a result of the failure of

the satellite position models to account for all forces
acting on the satellite. This can be eliminated
almost completely by using precise orbit data
(accurate within a few centimeters) that is made
available over the Internet
Contd..
Errors and
Biases

► Although a satellite’s atomic clocks is very stable, it

can still accumulate up to 17 ns of error each day,
which translates to a range error of 5 m. To correct
for this, the satellite clock is continually monitored
by the ground monitoring stations, and clock
corrections are transmitted periodically.
Contd..
Errors and
Biases

► The quality of the GPS location estimation depends

on how well the tracked satellites are spread across
the sky. In general, satellite geometry improves as
the distance between satellites increases. The size
of the uncertainty area (where the receiver could
be located) decreases as the distance between the
satellites increases”
Differential GPS

“Differential GPS takes advantage of the fact that a

satellite’s clock and the satellite coordinate errors, as
well as ionospheric and tropospheric delays exhibit high
temporal and spatial correlation and are very similar
even hundreds of kilometers apart
Contd
..

Differential techniques takes advantage of this by:

1. Coordinating multiple GPS receivers that
simultaneously track the same satellites
2. By having one or more GPS receivers fixed at
known positions, the observed errors from those
receivers can be transmitted to nearby roving
GPS receivers
3. These roving units are then able to reduce their
error in proportion to their proximity to the site
at which the correction was measured
Contd..
Real Time
DGPS

► Real-time differential GPS (DGPS) is a relative

positioning technique that provides sub-meter
accuracy
► A fixed receiver determines the DGPS corrections by
comparing its measured satellite ranges with ranges
computed using its known coordinates and the
satellite coordinates obtained from the navigation
message
Contd..
Real Time
DGPS

► The DGPS corrections are then transmitted over a

ground-based or a satellite-based wireless link to
the rover. The rover uses them to adjust its
ranging measurements”
Cellular Based Location
Systems
Based on Cell-ID

“A mobile-phone base station is typically equipped with

a number of directional antennas that define sectors of
coverage or cells each of which is assigned a unique cell
ID. Cell ID-based location is a simple technique where
the position of the mobile phone is estimated based on
the ID of the cell currently providing service to the
device”
Contd..
Based on Cell-ID

“Cell ID-based location is usually implemented on

the network side, and its key advantage are:
► It works for all phones, as no handset
modifications are needed.
► Accuracy depends on the size of the cell, ranging
from 150 m for micro-cells in urban cores to 30 km
for cells in rural settings”
Contd..
Based on Cell-ID

“Such accuracy may be sufficient for some applications

like weather or traffic report but falls short of the
requirements for another application like street
navigation, as such an application requires a cell phone
handset to be localized within 50m, 66% of the time”
Location Estimation with 802.11

“At their most basic, all of these systems work the same
way:
► 802.11 radios and their supporting drivers allow a
device to scan for nearby 802.11 APs
► Regardless of the variant of 802.11, or whether
encryption is enabled, these scans return a list of
the nearby APs and their unique IDs (called MAC
addresses)
► All of the 802.11 location systems capitalize on the
fact that
802.11 access points have limited range (typically
less than 100 m), and if a device can hear an
access point, it knows that it is in the vicinity of
that access point”
Contd..

“If a device can see more than one AP, it can estimate
its own location more precisely, and additional
information such as received signal strength and packet
loss rate can be further used to improve the accuracy of
location estimates”
Contd..
Why Use 802.11 for Location Estimation in Mobile Devices?

► Nearly all mobile devices have built-in 802.11

► There are many known mapped 802.11 access
points
► Higher densities increase the possibility of higher
coverage in the indoor/outdoor location without any
additional location infrastructure
Contd..
How do client devices discover and associate with the AP?

► The 802.11 protocol includes beacon frames that

an AP can send to alert clients of its presence
► The frequency with which these beacon frames are
sent commonly ranges from tens to hundreds of
beacon frames per second
► These frames contain the human readable SSID of
the network and the MAC address of the AP and
whether the AP is running any encryption or not
Contd..
How do client devices discover and associate with the AP?

► Client devices can passively learn about nearby

APs by listening for a small window of time on
each of the 802.11 channels. Alternatively, clients
also have the option of initiating an active scan by
sending a probe request which prompts access
points to reply with a probe response very similar
to a beacon frame
Contd..
Privacy Considerations

“The privacy of 802.11 location systems can be

excellent if it is implemented entirely with passive
radio reception and client-side computation.”
Location based Applications and Services

► Navigation and way-finding

► Asset tracking
► Emergency response
► Geofencing
► Location-based content and
search
► Social networking
► Health and wellness
► Gaming and entertainment

Exercise [In Class]: Discuss these use cases for location-

based appli- cation and services in detail.
Challenges in Location based Application and
Services

► Infrastructure and cost related

► Better algorithms and solutions
► User privacy
► Social understanding and
acceptance
► Others