Location collector platform for indoor tracking Ricardo Morales González and Miguel

Location collector platform for indoor tracking Ricardo Morales González  and Miguel
Master’s Thesis
Location collector platform for indoor
tracking
By
Ricardo Morales González and Miguel
Sanz Rodríguez
Department of Electrical and Information Technology
Faculty of Engineering, LTH, Lund University
SE-221 00 Lund, Sweden
Abstract
Human mobility patterns have become of great importance in different
topics like urban design as one of many examples in indoor environments.
The understanding of how human moves within an area remain limited due
to the lack of tools for monitoring the timing and location of individuals
and the necessity of user intervention.
Qubulus AB is an example of provider of an application position platform
for indoor tracking called LocLizard. LocLizard is based on a client
application installed on a user mobile phone. It collects radio downlink
signals to create a map with fingerprints that is used to calculate the
position of the mobile device. The analytics tool gives you the capability to
analyze historical data to see mobility patterns. However, the historical data
capable to be recorded is limited to the number of mobile devices, which
have Qubulus client software [1]. Other providers with similar solutions are
e.g. Polestar and Insiteo.
Our thesis project is to find a way to increase the number of mobile devices
that can be positioned within an indoor environment without a client
application installed on the mobile device. Furthermore, in order to reduce
user intervention, it is important to know if we can estimate the position of
the mobile device when it is in the idle mode state, i.e. when the mobile
device interacts with its backbone network on a M2M basis.
Our approach is to collect uplink signals from the Global System for
Mobile Communications (GSM), Universal Mobile Telecommunications
System (UMTS), and Wireless Local Area Networks (WLANs) that can be
used for positioning and tracking based upon location method, such as
trilateration or fingerprinting.
Three questions are to be answered: How often does the mobile device send
signals during the idle mode state for the different radio sources? Are these
uplink signals feasible for positioning and tracking? Is there a way to find
uniqueness between different mobile devices (terminals)?
Through our analysis and experimental results, trilateration as a location
method reports a positive performance.
2
Acknowledgments
There are many people who we would like to thank for their encouragement
and support in making our period of study a pleasant time. Here we can
only mention a few of them.
We would like to thank our supervisor, Henrik Hedlund, for his valuable
guidance in this master thesis.
We would like to thank the department of Electrical Information
Technology (EIT) of Lund Tekniska Högskola (LTH) and Professor Anders
Johansson for their helpful advice during this project.
We would like to thank Anders Berkeman for his wise advice to some key
aspects in this master thesis.
We would like to thank the department of Mathematical Statistics for their
contribution to this research.
And of course, thanks to our family and friends for their support during our
research within this master thesis.
Ricardo Morales González and Miguel Sanz Rodríguez
3
Table of Contents
Abstract ................................................................................................................. 2
Acknowledgments ............................................................................................... 3
Table of Contents ................................................................................................ 4
1
Introduction ............................................................................................... 6
1.1
1.2
1.3
1.4
1.5
2
Scope ............................................................................................................... 6
Problem statement ....................................................................................... 7
Research approach ...................................................................................... 8
Related work ................................................................................................. 9
Thesis organization ..................................................................................... 9
Theoretical background ..................................................................... 10
2.1 Indoor positioning techniques........................................................... 10
2.1.1 Fingerprinting ................................................................................................... 10
2.1.2 Trilateration ..................................................................................................... 11
2.2
Global System Mobile (GSM) ................................................................. 12
2.2.1 Overview ............................................................................................................ 12
2.2.2 Air interface and multiple access ................................................................ 12
2.2.3 Modulation ......................................................................................................... 13
2.2.4 Slow frequency hopping ................................................................................ 13
2.2.5 Uplink logical channels.................................................................................. 14
2.2.6 Frame hierarchy................................................................................................ 15
2.2.7 GSM burst and their types ............................................................................ 15
2.2.8 Idle mode ............................................................................................................ 16
2.3
Universal Mobile Telecommunication System (UMTS) ................... 21
2.3.1 Introduction to WCDMA .............................................................................. 21
2.3.2 UMTS system architecture ........................................................................... 22
2.3.3 UMTS radio interface ..................................................................................... 23
2.3.4 Physical structure ............................................................................................. 23
2.4
IEEE IEEE 802.11 ..................................................................................... 24
2.4.1 Overview ............................................................................................................ 24
2.4.2 IEEE IEEE 802.11b Physical layer ............................................................ 24
2.4.3 IEEE IEEE 802.11g Physical layer ............................................................ 26
2.4.4 DSSS-OFDM .................................................................................................... 28
2.4.5 Operating modes .............................................................................................. 28
2.4.6 IEEE IEEE 802.11 MAC protocol ............................................................. 29
2.4.7 Management frames ........................................................................................ 30
4
3
GSM/UMTS analysis.............................................................................. 37
3.1
3.2
3.3
3.4
4
Brief description of USRP ....................................................................... 37
Active mode ................................................................................................. 39
GSM and UMTS Attempt call ................................................................ 40
IDLE mode .................................................................................................. 43
Implementation ..................................................................................... 45
4.1
USRP approach .......................................................................................... 45
4.1.1 Detecting IEEE 802.11 signals using USRPs ....................................... 45
4.1.2 Scenario of measurements ......................................................................... 46
4.1.3 Procedure of measurements ..................................................................... 46
4.2
DD-WRT wireless broadband routers approach ............................... 51
4.2.1 Scenario of measurements ............................................................................ 52
4.2.2 Detecting IEEE 802.11 signals using DD-WRT wireless routers and
packet network analyzers .............................................................................................. 53
4.2.3 Limitations and assumptions ........................................................................ 54
4.2.4 Procedure of measurements .......................................................................... 54
5
Results....................................................................................................... 57
5.1
USRP results ............................................................................................... 57
5.1.1 Distinction between upstream and downstream ..................................... 57
5.1.2 Uniqueness of the signal ................................................................................ 58
5.2
Results of measurements with IEEE 802.11 wireless routers ......... 58
5.2.1 Idle mode experimental behavior ............................................................... 58
5.2.2 Variability analysis .......................................................................................... 61
5.2.3 Reproducibility analysis.............................................................................. 68
5.2.4 Practical application using trilateration ............................................... 72
6
Conclusions and future work ............................................................ 78
List of Figures ................................................................................................... 82
List of Tables..................................................................................................... 84
A.1 Python code .............................................................................................. 85
A.1 Parser code ............................................................................................... 90
A.1 LYNSYS WRT54GL specifications ...................................................... 94
A.1 Getting started with dd-wrt ................................................................ 96
A.1 Code to install tcpdump on DD-WRT routers ............................... 99
A.1 Getting started with the USRP..........................................................100
5
CHAPTER
1
1
Introduction
1.1 Scope
The necessity of knowing the location of people is increasing day by day; at
the same time, technology continues revolutionizing our world. Location
Based Services (LBS) have acquired a huge importance for both mobile
operators and users, since they posses great commercial opportunities but
also provide great utility for users in daily life. The GPS (Global
Positioning System), that a few years ago came up as an innovative
technology, is nowadays a widespread service. GPS technology has proven
successful technology for outdoor environments, but no technology has yet
gained a similar wide-scale adoption for indoor positioning. This is a new
exciting research scope that has just started to develop, but is evolving
quickly as the demand of indoor localization increases.
There are plenty of uses of indoor localization. Airports, malls, and
property owners are just a few examples of organizations that can get
benefits using indoor positioning. For instance, the airports are nowadays
looking for ways to raise customer satisfaction in a very competitive
environment. To give the traveler the possibility to position themselves
accurately opens up for services where they can get accurate calculations on
how long it will take them to get to the gate combined with immediate
information on departure & check in time, potential delays and much more.
For malls and property owners’ indoor localization brings them the
technical possibilities to meet their customers all the way from approaching
their location, entering and moving around indoor. Combining store and
product search with event marketing, discrete advertising, offers and
coupons, all controlled by the position of the customers, means definitely a
huge commercial potential [2].
There exist many different positioning methods that given suitable
measurements can be used to estimate sensor positions. In recent years,
positioning systems for indoor areas using the existing wireless networks
such as GSM, UMTS and local area network (WLAN) infrastructure have
6
been developed. Remarkable techniques are angle-of-arrival (AOA), timeof-arrival (TOA), multilateration or location fingerprinting. Fingerprinting
method uses the received signal strength (RSS) that the mobile device
receives from wireless networks at the sampling locations to build a "radio
map" for the target environment. Once the radio map is built, the position is
computed by comparing a new fingerprint with the ones forming the radio
map, using different algorithms. Location fingerprinting performs well for
non-line-of-sight (NLOS) circumstances and LOS environments are not
required. It is generally agreed that a desirable indoor location system
should be characterized by high accuracy, short training phase, costeffectiveness (preferably using off-the-shelf hardware), and robustness in
the face of previously unobserved conditions. Such schemes can provide
value-added services for existing WLANs and 3G networks. Indoor
localization techniques using location fingerprints are gaining popularity
because of their cost-effectiveness compared to other infrastructure-based
location systems. The fingerprinting technique is relatively simple to
deploy compared to the other techniques such as angle-of-arrival (AOA)
and time-of-arrival (TOA). Moreover, there is no specialized hardware
required at the mobile station (MS). Any existing wireless LAN or 3G
infrastructure can be reused for this positioning system [3][4].
Qubulus is one of the most promising indoor positioning start-ups, and it
bases its technology in location fingerprinting. They use RSSI, from
existing radio access networks such as GSM, UMTS and WLAN, at mobile
devices to create the radio map that enables to compute the 3D position
when a user requires so. This method gives ~ 3 meter accuracy in X and Y
axis, and ~ 1 meter in Z axis. Currently the technology is based on a free
client application installed in an android mobile device, in which users can
see their current position [2].
There are other companies like GloPos, which has a similar technology
collecting radio signals from different sources and applying some
algorithms for the positioning of the mobile phone in an indoor
environment.
1.2
Problem statement
The current techniques of indoor localization are based on signals received
in the downlink (from Base Station or Access Point to the mobile device).
For this reason, they require a client application installed on the mobile
7
devices, and therefore active user intervention is essential.
An increasing important demand of knowing mobility patterns of people in
indoor scenarios has turned into a great commercial potential. For this
purpose, to have client applications installed on every terminal and to need
that the user intervenes actively are a clear obstacle. To develop a system
that makes possible to detect and position mobile devices without any
application installed on the terminals and without any (or minimum) active
user intervention is the challenge that this master thesis faces.
1.3 Research approach
An alternative and novel approach for indoor positioning is developed in
this master thesis. Instead of using RSS from downlink signals, we detect
uplink GSM, UMTS and Wi-Fi signals coming out from mobile devices in
order to get power levels that can be used for computing position using any
indoor localization technique such as fingerprinting or trilateration. We
focus our research in the idle mode of the three quoted technologies since is
the most common scenario. Since this is a novel approach, the challenge is
to prove the feasibility of uplink signals to be used for computing indoor
position.
For the purpose of analyzing UMTS and GSM signals we combined
Software defined radio (SDR) GNU Radio with Universal Software Radio
Peripheral (USRP) equipment. These tools are used to design and
implement a wireless spectrum sensor that captures traffic from mobile
devices and store it in files that are further processed in Matlab.
In order to analyze Wi-Fi signals two approaches that differ on the way that
the signals are detected are employed. Firstly, we utilized GNU Radio and
USRP to detect and characterize uplink Wi-Fi signals. Secondly, LinkSys
WRT54GL routers are used to sense the IEEE 802.11 spectrum and detect
uplink packets, together with packet network analyzers such as Wireshark
and tcpdump. The firmware DD-WRT, which is a Linux based alternative
OpenSource firmware suitable for a great variety of WLAN routers and
embedded systems, is installed on the routers in order to provide additional
capabilities.
In all tests and measurements, two different mobile smartphones (HTC
Desire Z and Samsung Wave GT-S8500) were used in order to analyze the
8
possible differences between vendors.
1.4 Related work
As this is a novel approach to indoor positioning, we have not found any
related work similar to our master thesis. Obviously we have used plenty of
bibliography and references from academic papers and dissertations that
have been relevant to our research, but it is worth to mention that this
master thesis presents a novel and innovative method for indoor
localization that nobody has researched into before.
1.5 Thesis organization
The outline of the whole dissertation is presented below:

Chapter 2 explains the theoretical background needed to understand
this master thesis. It provides an overview of the three radio access
technologies employed in the project and a detailed explanation of
the idle mode in each of these technologies.

Chapter 3 presents the analysis of UMTS and GSM uplink signals
using GNU Radio and USRP equipment. The feasibility of these
signals for the purpose of indoor positioning is discussed here.

Chapter 4 describes the implementation of two different approaches
for the purpose of detecting IEEE 802.11 uplink signals. For every
approach, the method of detection, scenario and procedure of
measurements are explained in detail.

Chapter 5 provides experimental results of each approach with the
corresponding explanations. For the LinkSys routers approach, an
example of the use of the results to compute specific indoor
positioning is presented, using trilateration technique.

Chapter 6 summarizes the conclusions obtained after the work
within this master thesis and future work.
9
CHAPTER
2
2
Theoretical background
2.1 Indoor positioning techniques
Indoor positioning has become an important research topic for industrial
and academic purposes. Accurate localization in an indoor environment can
enable localization aware services in many domains. Many domains can get
benefits from indoor localization services. Furthermore, localization
information of mobile units can be recorded to provide historical data of
mobility patterns. Hence, the historical data recorded can be used for better
designs, distribution of infrastructure among others regarding the domain.
There are many position techniques that have been proposed; however, not
all techniques are suitable for indoor environments due to multipath effect.
Fingerprinting is one that could overcome the multipath drawback.
2.1.1 Fingerprinting
Fingerprinting consists of two stages: A training phase where the signal
strengths from different radio sources are collected (like GSM, UMTS, and
Wi-Fi) at reference points and store together with their physical coordinates
on a server; they are called fingerprints. In the positioning determination
phase, the mobile device performs signal strength measurements and is
continuously uploaded to the application back-end in real time, or at a
configurable rate of frequency. The signal strengths collected are matched
for similar patterns with different algorithm in the database. The closest
match is selected and returned as the estimated position[2].
The signal strength is sensitive to the environment. Thus, a collection of
access points can be chosen for fingerprinting location. The following
factors can affect the signal strength [5].
10
1. Temperature: it affects the signal strength and can vary the
fingerprints.
2. Sampling time: signal strength may differ from the same location in
sampling time.
3. Granularity: The distance between two adjacent locations, where
signals strengths are collected, affects the accuracy of location
sensing. For our case, it was taken 5 meters away.
4. Pattern algorithm recognition: the most important part in location
fingerprinting to compare multiple fingerprints stored in the server
and picks the one that best matches the observed RSS.
2.1.2 Trilateration
The uplink data collected in this thesis project will be given to the
Mathematical Statistics Department at Lund University where it will be
analyzed to create algorithms for indoor positioning techniques.
Why trilateration? As an alternative of the fingerprinting technique,
trilateration was used in a simple way to estimate the position.
Trilateration is the process to calculate the position of points by
measurement distances. Trilateration has practical applications in surveying
and navigation, including Global Positioning Systems (GPS).
Trilateration can be either on two or three dimensions. The difference is that
we have two circles in two dimensions and spheres in three dimensions.
When a point lies on two circles, the circle centers and the radii provide
enough information to narrow the possibilities locations down to two. [6]
In three dimensional, when the point now lays on three circles, the centers
of the circle as well their radii down the possibilities to no more than two. If
it is increased the number of circles, it can be reduce the possibilities of
locations.
The intersection of the surfaces for the three spheres is found by
formulating the functions of them.
11
By resolving the equations (1), (2), and (3) it can be obtained x, y and z and
get the location of the point.
2.2 Global System Mobile (GSM)
2.2.1 Overview
GSM is a digital wireless network standard designed by standardization
committees from major European telecommunications operators and
manufacturers. The GSM standard provides a common set of compatible
services and capabilities to all mobile users across Europe and several
million customers worldwide [7]. In this section we will describe the GSM
features that are relevant to our research.
2.2.2 Air interface and multiple access
GSM employs a combined Frequency Division Multiple Access
(FDMA)/Time Division Multiple Access (TDMA) approach which further
combines with Frequency Division Duplex (FDD). In GSM 900, the most
spread version of GSM, FDD is used to multiplex the uplink (890-915
MHz) and the downlink (935-960 MHz). Frequencies always exist in pairs,
one for uplink (Mobile Station to Base Station) and one for downlink (Base
Station to Mobile Station). The frequency spacing between the uplink and
downlink for any given connection is 45 MHz. However, in the case of
certain channels called “common channels” it is possible that the uplink and
downlink frequencies serve different purposes and may not have any
relation with each other. In contrast, channels that are dedicated for a
Mobile Station (MS) are always allocated in pairs [8].
The air interface of GSM uses FDMA to operate over various frequencies
separated 200 KHz. The outer 100 KHz of each 25 MHz band are not used,
as they are guard bands to limit interference in the adjoined spectrum which
12
is used by other systems. The remaining 124 duplex channels are numbered
by the so-called Absolute Radio frequency Channel Numbers (ARFCNs). In
each of these sub bands, GSM performs TDMA to split the time axis in 8
time slots, which are periodically available to each of the possible 8 users.
The collection of these 8 time slots repeated periodically over time is called
a TDMA frame. The TDMA frames on the uplink are transmitted with a
delay of 3 time slots with regard to the downlink. Each time slot that recurs
in every TDMA frame and operated over a specific frequency is viewed as
the `physical channel´ that carries user/signaling information. This physical
channel lasts for 15/26 ms or 576.9 microseconds. The information is
carried at a modulation rate of 270,833 Kbps and it is equivalent to 156.25
bits. Within each frame the timeslots are numbered from 0 to 7, and each
subscriber accesses one specific timeslot in every frame on one frequency
sub band. A MS uses the same time slots in the downlink as in the uplink,
i.e. the time slots with the same number. Owing to the shift of three time
slots, a MS does not have to send as the same time as it receives, and
therefore does not need a duplex unit [8][9].
2.2.3 Modulation
The modulation used in GSM is Gaussian Minimum Shift Key (GMSK),
which is a continuous-phase frequency-shift keying modulation scheme. It
is similar to standard minimum-shift keying mobile (MSK); however the
digital data stream is first shaped with a Gaussian filter before being
applied to a frequency modulator. Special advantages of this modulation
scheme are the narrow transmitter power spectrum with low adjacent
channel interference and a constant amplitude envelope which allows the
use of simple amplifiers in the transmitter with special linearity
requirements [8].
2.2.4
Slow frequency hopping
Mobile radio channels suffer from frequency selective interferences, e.g.
frequency selective fading due to multipath propagation phenomena. This
selective frequency interference can increase with the distance from the
base station, especially at the cell boundaries and under unfavorable
conditions. Frequency hopping changes the transmission frequencies
periodically and thus average the interference over the frequencies in one
cell. This improves the Signal to Noise Ratio (SNR) to a high enough level
for good speech quality [10].
13
GSM provides a slow frequency hopping which changes to a different
frequency with each burst. The scheme is based on the idea that every
mobile station transmits its TDMA frames according to a sequence of
frequencies specified by the frequency hopping algorithm. A mobile station
transmits on a fixed frequency over one time slot (577 microseconds) and
then jumps to another frequency before the next TDMA frame. There are
64 frequency hopping sequences in GSM, each one containing up to 64
frequencies. The result hopping rate is about 217 changes per second,
corresponding to the TDMA frame duration [10].
2.2.5
Uplink logical channels
In GSM the channels are divided into two groups: physical and logical
channels. A physical channel corresponds to a specific timeslot on one
carrier, while a logical channel reflects the specific type of information that
the physical channel carries. The logical channels are mapped to the
physical channels, and they carry different information depending on the
type of logical channel.
The logical channels are divided into traffic and control channels. The
traffic channels are the resources available for the users to send data or
voice. The control channels are used for signaling and controlling the traffic
channels [10]. Below the uplink logical channels are shortlisted [10]:
The Random Access Channel (RACH): this channel is used initially by
the MS to attempt accessing the network. When the MS responds to a
paging request or when initiating a call, the MS is requesting a signaling
channel.
The Stand-alone Dedicated Control Channel (SDCCH): this bidirectional channel carries all the signaling information. It can be used for
authentication, ciphering, call set-up or transmission of text messages.
The Slow Associated Control Channel (SACCH): this bi-directional
channel is used to transfer signaling information during an ongoing call on
a traffic channel. In the uplink the MS reports the downlink measurements
to the BTS. It is used for non-urgent procedures, and can carry up to two
messages per second in each direction.
14
The Fast Associated Control Channel (FACCH): this bi-directional
channel is used when there is a need of higher capacity signaling in parallel
with ongoing traffic, and it is mainly used for handover procedures.
2.2.6 Frame hierarchy
The physical content of each time slot is called a `burst´. There are different
types of bursts, each of which has a different structure. Each burst maps to
a time slot, and they form the lowest level in GSM lever hierarchy. The
TDMA frame thus forms the next level in the hierarchy. At the next level
there is a collection of TDMA frames called the `multi-frame´ that is of 2
types, 26-frame multi frame and 51-frame multiframe. The 26-frame
multiframe is used for traffic channels wherein 24 channels are used for
actual traffic exchange, while one is used for low-rate signaling that
accompanies any traffic exchange, and one channel is left idle. In contrast,
the 51-frame multiframe is used for signaling. The difference at the multiframe level is removed at the `superframe´ level wherein a superframe
always has 1326 frames (26x51 or 51x26). The last level of the hierarchy is
the `hyperframe´ which comprises 2048 frames. Thus, the GSM hierarchy
repeats itself after an interval of 3 hours 28 minutes 53 seconds 760
miliseconds [9].
2.2.7 GSM burst and their types
A burst lasts for 156.25 bit duration or a period of 576.9 microseconds.
During this period, the signal reaches from amplitude of 0 to a specified
value. Thereafter, the signal is modulated and transmitted. At the end of the
transmission, the signal reaches back to 0. The transmitted signal here is
viewed to be the modulated signal obtained after modulating a signal that
comprises a series of 1´s followed by the information to be sent which is
followed again by a series of 1´s. In order to ensure efficient demodulation,
3 tail bits (all 0´s) are added before and after the information contents to
ensure the transition of the signal (from 1 to 0 and back from 0 to 1) during
the transfer. The guard period is used to ramp up and down the signal [9].
There are different types of bursts sent in the uplink [9]:
Normal burst (NB): this burst is used to carry information on traffic and
control channels, except for RACH (Random Access Channel). It contains
15
116 encrypted symbols and includes a guard time of 8.25 symbol duration
(30.46 μs).
Access burst (AB): this burst is used in order to help the mobile to find its
distance. Its useful contents are much shorter than ones of the normal bursts
to allow for propagation delays8 only 36 information bits can be carried).
When the mobile first attempts to access the network, it does not provide its
own identity; instead, it seeks a channel on which it can send further
information needed for authentication and resource allocation. There is no
training sequence in this burst for reducing signaling complexity, but a
fixed synchronization pattern is used. The access burst is used to determine
the distance of the mobile from the base station. This scenario arises in two
basic cases. Firstly, this happens when the mobile makes the first attempt
for signaling (e.g. Location Update) or for communication (e.g. mobile
originated call or SMS). The second scenario occurs when the mobile
moves from one cell to another (e.g. handover). In this case the first burst to
the new base station is the access burst and it is used to synchronize the
mobile with the new base station.
Higher symbol rate burst (HB): this burst is used to carry information on
full rate packet data traffic channels using higher symbol rate. It contains
138 encrypted symbols and includes a guard time of 10.5 reduced symbol
periods.
2.2.8 Idle mode
The idle mode behavior is managed by the MS. It can be controlled by
parameters which the MS receives from the base station on the Broadcast
Control Channel (BCCH). All of the main controlling parameters for idle
mode behavior are transmitted on the BCCH carrier in each cell. The idle
mode tasks can be divided into 4 processes: Public Land Mobile Network
(PLMN) selection, cell selection, cell re-selection, and location updating.
PLMN selection is the process in which the MS tries to connect or associate
with a Public Network after being powered on or after experimenting lack
of coverage [11]. In this section we will not explain in detail PLMN
selection, since the cases where it takes place are not within our area of
interest because either they require active user intervention or they occur in
very special conditions, but not normal conditions when a person moves in
an indoor scenario.
16
2.2.8.1 Cell selection
The cell selection algorithm tries to find the most suitable cell of the
selected PLMN according to various requirements. If no suitable cell is
found and all available and permitted PLMNs have been tried, the MS will
try to camp on a cell irrespective of PLMN identity and enter a limited
service state. In this state the MS will be able to make emergency calls
only. If the MS loses coverage it will return to the PLMN selection state
and select another PLMN. There are two different strategies to perform cell
selection: normal cell selection or stored cell list selection [11]. In this
section we will explain normal cell selection since it is the most used.
The choice of such a suitable cell for the purpose of receiving normal
service is referred to as "normal camping". There are various requirements
that a cell must satisfy before an MS can perform normal camping on it
[11]:
1) It should be a cell of the selected PLMN or, if the selected PLMN is
equal to the last registered PLMN, an equivalent PLMN.
2) It is not barred (when a cell is barred it will not be camped on by an
MS in idle mode but a MS in dedicated mode can perform handover
to it).
3) It does not belong to a location area included in the list of
“forbidden location areas for roaming”. The location areas that are
forbidden, after an attempt to do a location updating has failed, will
be stored in the MS as forbidden location areas for national
roaming. This list will be cleared when the mobile is powered off.
4) The radio path loss between MS and BTS must be below a threshold
set by the PLMN operator.
5) It should not be a SoLSA (Cells on which normal camping is
allowed only for MS with Localized Service Area subscription)
exclusive cell to which MS does not subscribe. This requirement is
only valid for MSs supporting SoLSA.
Initially, the MS looks for a cell which satisfies these 5 constraints
("suitable cell") by checking cells in descending order of received signal
17
strength. If a suitable cell is found, the MS camps on it and performs any
registration necessary. Cells can have two levels of priority; suitable cells
which are of low priority are only camped on if there are no other suitable
cells of normal priority.
In order to speed up these processes, a list of the RF channels containing
BCCH carriers of the same PLMN is broadcast in the system information
messages. Also, the MS does not need to search all possible RF channels to
find a suitable cell [11].
When the MS has no information on which BCCH frequencies that are used
in the network, the MS will search all RF channels in its supported
frequency band, take measurement samples of the received RF signal
strength and calculate the received average level for each. The average is
based on at least five samples per RF carrier spread evenly over a 3 to 5
second period. The MS searches at least the number of the strongest RF
channels in descending order of RSS to see which ones are BCCH carriers.
If no BCCH carriers have yet been found, searching will continue until at
least one BCCH carrier is found. The first BCCH carrier found which is
from a suitable cell and on which there is a normal priority indication is
taken and that cell is camped on. If at least the numbers of the strongest RF
channels have been tried and the only suitable cells found have low priority
indication the MS shall camp on the strongest of these cells. If at least the
30 strongest GSM 800 or GSM 900 RF channels or 40 strongest GSM 1800
RF channels or 40 strongest GSM 1900 RF channels have been tried and no
suitable cell was found, the MS will select another PLMN according to the
PLMN selection procedure and search for suitable cells there [11].
2.2.8.2 Cell re-selection
After a cell has been successfully selected, the MS will start the cell
reselection tasks. It will continuously make measurements on its current
serving cell and neighboring cells, in order to initiate cell reselection if
necessary. The following events trigger a cell reselection [11]:




The path loss to the cell has become too high.
There is a downlink signaling failure.
The cell camped on (current serving cell) has become barred.
There is a better cell in terms of the path loss criterion in the
same registration area, or a much better cell in another
registration area of an equivalent PLMN.
18

Upper layers have determined that the network has failed an
authentication check.
The MS continuously monitors all neighboring BCCH carriers, as indicated
by the BA list, in addition to the BCCH carrier of the serving cell, to detect
if it is more suitable to camp on another cell. At least five received signal
level measurement samples are required for each defined neighboring cell.
A running average of the received signal level will be maintained for each
carrier in the BA list. All system information messages sent on BCCH must
be read at least once every 30 seconds in order to monitor changes in cell
parameters. The MS also tries to synchronize to and read the BCCH
information that contains parameters affecting cell reselection for the six
strongest non-serving carriers (in the BA list) at least every five minutes.
The MS also attempts to decode the BSIC parameter for each of the six
strongest cells, to confirm that it is still monitoring the same cells [11].
Before camping on the cell after re-selection, the MS shall attempt to
decode the full set of system information. The MS shall check that the
parameters affecting cell re-selection are unchanged. If a change is detected
the MS shall check if the cell re-selection criterion is still valid using the
changed parameters. If the cell selection criteria are still valid, the MS shall
camp on the cell. If they are not still valid, the MS shall repeat this process
for the cell with the next highest RSS [11].
2.2.8.3 Location updating
To make it possible for the mobile subscriber to receive a call, the network
must know where the MS is located. To keep the network updated on the
location of the MS, the system is informed by the MS on a regular basis.
This is called Location Updating. There are three different types of location
updating defined: normal, periodic registration and International Mobile
Subscriber Identity (IMSI) attach/detach [12].
Normal location updating: the location update is initiated by the MS when
it detects that it has entered a new location area. The MS listens to the
system information, compares the Location Area Identity (LAI) to the one
stored in the MS on the SIM card (on BCCH channel if idle or SACCH
channel if active) and detects whether it has entered a new location area or
is still in the same location area. If the broadcast LAI differs from the one
19
stored on the SIM card, the MS must perform a normal location update
following the next steps. [12]
1) The MS sends a channel request message including the reason
for the access. Reasons other than location updating can be for
example, answering a page or emergency call.
2) The message received by the BTS is forwarded to the BSC. The
BSC allocates an SDCCH, if there is one idle, and tells the BTS
to activate it.
3) The MS is now told to tune to the SDCCH.
4) The MS sends a location updating request message that contains
the identity of the MS, the identity of the old location area and
the type of updating.
5) The authentication parameter is sent to the MS. In this case the
MS is already registered in this MSC/VLR and the
authentication parameter used is stored in the VLR. (If the MS is
not already registered in this MSC/VLR the appropriate HLR or
the previously used MSC/VLR must be contacted to retrieve MS
subscriber data and authentication parameters).
6) MS sends an answer calculated using the received authentication
parameter.
7) If the authentication is successful, the VLR is updated. If
needed, the HLR and old VLR are also updated.
8) The MS receives an acceptance of the location updating.
9) The BTS is told to release the SDCCH.
10) The MS is told to release the SDCCH and switches to idle mode
Periodic Registration location updating: to reduce unnecessary paging of
a mobile that has left the coverage area, has run out of battery power or for
any other reason has the wrong status in the Mobile Switching Center
(MSC)/Visitor Location Register (VLR), there is a type of location
20
updating called periodic registration. The procedure is described below
[12]:
1) MS listens on the BCCH to specify if Periodic Registration
Location Update is used in the cell. If periodic registration is used,
the MS is told how often it must register. The time is set by the
operator and can have values from 0 to 255 deci-hours (a unit of six
minutes). If the parameter is equal to zero, periodic registration is
not used in this cell. If the parameter is set to ten, for example, the
MS must register every hour.
2) Both the MS and the MSC have the timer which controls the
procedure. When the timer in the MS expires, the MS performs a
location updating, type periodic registration. After that, the timers in
MS and MSC restart. In the MSC there is a time scanning function
for the MSs. If the MS does not register within the determined
interval plus a guard time, then the scanning function in MSC
detects this and the MS is flagged as detached.
This process is controlled by the T3212 parameter, which is a
timeout value broadcast to the MS in the system information
messages. The interval ranges between six minutes (T3212 = 1) and
25.5 hours (T3212 = 255). The periodic registration timer is
implemented in the MS, and it will be reinitiated every time the MS
returns to idle mode after being in dedicated mode [12]
IMSI attach/detach: The IMSI attach/detach operation is an action taken
by an MS to indicate to the network that it has entered into idle
mode/inactive state. When an MS is powered on, an IMSI attach message is
sent to the MSC/VLR. When an MS is powered off, an IMSI detach
message is sent. Therefore, it requires active user intervention and it is not
relevant for our approach [12].
2.3 Universal Mobile Telecommunication System (UMTS)
2.3.1 Introduction to WCDMA
WCDMA was chosen as the radio technology for universal mobile
telecommunication system (UMTS) as is shown in fig. 1. There are two
modes of operation frequency division duplex (FDD) and time division
21
duplex (TDD). In the FDD, the uplink and downlink uses different
frequency bands meanwhile in the TDD, the uplink and downlink share a
frequency band by using synchronize intervals. Our work focuses on
UTRA-FDD uplink as it is mainly used in Europe and we want to analyze
the signal quality in the reverse link [13].
2.3.2 UMTS system architecture
The three main components of UMTS is the user equipment, the UMTS
terrestrial radio-access network (UTRAN), and the core network.
Furthermore, there are two interfaces, the Iu interface between the core
network and the UTRA network, and the Uu radio interface between the
UTRA network and the user equipment. The UTRAN comprises some of
the following functions: encryption/decryption, power control, modulation,
multiplexing, etc [13].
Fig. 1. Components of the UMTS system
22
2.3.3 UMTS radio interface
WCDMA uses a basic chip rate of 4.096 Mchip/s and different spreading
factors. The different rates that UMTS can support are due to the fact of
different spreading factors. As it is implied by UTRAN-FDD, uplink and
downlink use different frequencies. For Europe, the uplink uses the
frequencies between 1920 and 1980 MHz and the base station uses the
frequencies between 2110 to 2170 MHz.
2.3.4 Physical structure
There are two types of dedicated channels: the dedicated physical data
channel (DPDCH), which carries data generated at layer two or above; and
the dedicated physical control channel (DPCCH), which carries control
information at layer 1. Each connection has one DPCCH and one or many
DPDCH’s.
There are some common defined channels: primary and secondary common
control physical channels (CCPCH). They carry downlink information,
synchronization channel (SCH), which are used for cell search and the
physical random access channel (RACH) [14].
Fig. 2 shows the structure of a UTRA-FDD (WCDMA) frame. It can be
seen on the figure that a radio frame comprises 15 time slots of length 10
ms. Time slots are used to support periodic functions whereas GSM slots
are used for user separation. Each time slot contains 2,560 chips. The
bandwidth per WCDMA channel is 4.4 to 5 MHz. The DPDCH carries
layer 2 data, while the DPCCH carries pilot bits, transmit-power control
(TPC) commands, the feedback information field (FBI) and the transport
format combination identifier (TFCI) [9].
The DPCCH and DPDCH are mapped to I and Q components and spread to
the chip rate with two different channelization codes. The resulting
complex signal is scrambled and QPSK modulation with root-raised cosine
pulse shaping with rolloff factor of 0.22 is applied in the frequency domain.
Orthogonal variable spreading factor (OVSF) codes are used for the
channelization which they are used to preserve orthogonality between
physical channels with different rates and spreading factors. Channelization
codes and UE-specific scrambling codes are given by the network.
Idle mode and uplink channels are similar to GSM.
23
Fig. 2. Uplink frame structure
2.4 IEEE IEEE 802.11
2.4.1 Overview
The IEEE 802.11 family consists of a series of over-theair modulation techniques that use the same basic protocol. Wireless Local
Area Networks (WLAN) based on IEEE 802.11 are been deployed with a
great success on a great variety of home, office and corporate
environments. Since the introduction of the IEEE 802.11 standard, multiple
extensions and amendments have been proposed and accepted, but the most
popular and worldwide spread are those defined by the IEEE 802.11b and
IEEE 802.11g protocols, and therefore they will be treated in this section.
Both standards work in the band of 2.4 GHz, and they both occupy 13
channels separated 5 MHz with a bandwidth of 22 MHz. The different
channels can be seen in table 1.
2.4.2 IEEE IEEE 802.11b Physical layer
IEEE IEEE 802.11 uses two different spread spectrum techniques in its
physical layer, which are incompatible:
Direct-Sequence Spread Spectrum (DSSS): DSSS spreads the data across
a broad range of frequencies using a high-rate code sequence with binary
phase shift keying (BPSK) to modulate the carrier for spreading.
24
TABLE 1. LIST OF IEEE 802.11 B/G CHANNELS [13]
Channel
Center Frequency
Channel Width
(GHz)
(GHz)
1
2.412
2.401 - 2.423
2,3,4,5
2
2.417
2.406 - 2.428
1,3,4,5,6
3
2.422
2.411 - 2.433
1,2,4,5,6,7
4
2.427
2.416 - 2.438
1,2,3,5,6,7,8
5
2.432
2.421 - 2.443
1,2,3,4,6,7,8,9
6
2.437
2.426 - 2.448
2,3,4,5,7,8,9,10
7
2.442
2.431 - 2.453
3,4,5,6,8,9,10,11
8
2.447
2.436 - 2.458
4,5,6,7,9,10,11,12
9
2.452
2.441 - 2.463
5,6,7,8,10,11,12,13
10
2.457
2.446 - 2.468
6,7,8,9,11,12,13
11
2.462
2.451 - 2.473
7,8,9,10,12,13
12
2.467
2.456 - 2.478
8,9,10,11,13
13
2.472
2.461 – 2.483
9,10,11,12
25
Overlapped
Channels
This is combined with lower rate data that also modulates the carrier with
either differential BPSK (DBPSK) or differential quadrature PSK
(DQPSK). The high-rate code is an 11-bit Barker code that provides 11
MHz chipping rate and has good autocorrelation properties and gives
waveform protection to interference and multipath. The data rate varies
between 1, 2, 5.5 and 11 Mb/s (DBPSK and DQPSK modulation). The
typical receiver can operate with a 0 dB signal-to-noise ratio in the
spread bandwidth and therefore can tolerate strong multipath and delay
spread. DSSS uses a lower power density (power/frequency), making it
harder to detect. DSSS also sends redundant copies of the encoded data to
ensure reception. Narrowband interference appears to the receiver as
another narrowband transmission. When the total received signal is
decoded, the wider band transmission (DSSS encoded data) is decoded with
the same code back to its original narrowband format while the interference
is decoded to a lower power density signal, thereby reducing its effects.
When broadband interference is present, however, the resulting decoded
broadband interference can give a much higher noise floor, almost as high
as the decoded signal. For this reason, DSSS works better for large data
packets in a low to medium interference environment, but not as well in
higher interference industrial applications [15].
Frequency Hopping Spread Spectrum (FHSS): in this method the 2.4
GHz band is divided into 79 channels of 1MHz, each one modulated with
Gaussian Frequency Shift Modulation (GFSK). Sender and receiver match
the hop pattern of switching the channel and data is transmitted in different
channels according to the frequency pattern. Each data transfer in
implemented according to a different hop pattern, and the hop patterns are
developed so as to minimize the probability of usage of the same channel
simultaneously by two users. FHSS is a robust technology, with good
behavior in harsh environment characterized by large areas of coverage,
multiple collocated cells, noises, multipath or Bluetooth presence, but it is
limited to just 2 Mb/s data rate, which makes it less popular than DSSS
[16].
2.4.3 IEEE IEEE 802.11g Physical layer
IEEE
IEEE 802.11g uses Orthogonal
Frequency Division
Multiplexing (OFDM), a Frequency-Division Multiplexing (FDM) scheme
used as a digital multi-carrier modulation method. A large number of
closely-spaced and overlapped orthogonal sub-carriers are used to
26
carry data. By carefully specifying the symbol duration and frequency
spacing, subcarriers will generate no interference or distortion on adjacent
subcarriers. The data is Viterbi encoded prior to transmission and
distributed across several parallel data streams carrying by 52 sub-carriers
via an interleaver. Each sub-carrier is modulated with a conventional
modulation scheme (such as QAM or PSK) at a low symbol rate,
maintaining total data rates similar to conventional singlecarrier modulation schemes in the same bandwidth. After modulation, N
parallel symbol streams (at 1/N of the original rate) are fed to the N point
Inverse Fast Fourier Transformer (IFFT). After IFFT, a cyclic prefix (CP) is
added at the beginning of the symbol to make the signal robust against
multipath propagation [17].
The IEEE 802.11g OFDM symbol rate is 250 KHz, corresponding to a
symbol period of 4 microns. The OFDM pulse contains a guard interval,
which is 800ns in duration. During the guard interval, inter-symbol
interference (ISI) from a delayed pulse arriving via a delayed path may
occur. However, once the signal arrives at the receiver and is translated to
the digital domain, the guard interval is essentially discarded by the
baseband processor. With the guard interval eliminated, a 3.2s rectangular
FFT window remains. The zero-ISI condition in the frequency domain is
met due to the use of a 3.2s FFT window in conjunction with 312.5 KHz
subcarrier spacing. In this manner, cross-talk among subcarriers is
eliminated [17].
IEEE IEEE 802.11 g OFDM offers a variety of data rates which covers 6, 9,
12, 18, 24, 36, 48, and 54 Mbit/s, depending on the modulation and coding
rate as shown on table 2.
27
TABLE 2. DATA RATES IN IEEE 802.11G
Data
Rate
(Mb/s)
Modulation
Coding
rate
6
9
12
18
24
36
48
54
BPSK
BPSK
QPSK
QPSK
16-QAM
16-QAM
64-QAM
64-QAM
½
¾
½
¾
½
¾
2/3
¾
Coded
bits per
subcarrier
1
1
2
2
4
4
8
8
Coded
bits per
OFDM
symbol
48
48
96
96
192
192
288
288
Data bits
per
OFDM
symbol
24
36
48
72
96
144
192
216
2.4.4 DSSS-OFDM
Because of the quick success of IEEE 802.11g, the IEEE implemented a
mode where IEEE 802.11b and IEEE 802.11g could coexist in the same
network. With this functionality, an IEEE 802.11g network can admit IEEE
802.11b terminals and vice versa. This hybrid modulation has the DSSS
preamble and header and the OFDM payload. Although both standards can
coexist in the same network and their hardware is fully compatible, the
overall performance decreases significantly in a mixed network. Whereas
IEEE IEEE 802.11b reaches up to 11 Mb/s, IEEE IEEE 802.11g can reach
up to 54 Mb/s, but its speed decreases considerably when there exists
legacy IEEE 802.11b participants in the network [18].
2.4.5 Operating modes
Infrastructure mode: in this mode a set of wireless terminals are
connected to an access point (AP) forming the wireless network and the AP
is connected to a wired network. Such combinations are called Basic
Service Set (BSS), and two or more BSS form an Extended Service Set
(ESS). Since most of the wireless stations need to access a server, a printer
or to Internet available in a wired LAN, this mode is the most used [16].
28
Ad-Hoc: in this mode the wireless stations are connected directly to each
other forming a simple network, without the intervention of an AP [16].
2.4.6 IEEE IEEE 802.11 MAC protocol
The IEEE IEEE 802.11 defines two basics methods to access the medium:
Distributed Coordination Function (DCF): The fundamental access
method of the IEEE IEEE 802.11 is a DCF known as a Carrier Sense
Multiple Access with Collision Avoidance (CSMA/CA). For a station
(STA) to transmit, it shall sense the medium to determine if another STA is
transmitting. This is doing by detecting a minimum duration called DCF
Interframe Space (DIFS) to see if there is any other transmission in
progress on the wireless medium. Afterwards, according to the current
Contention Window (CW), the station keeps further sensing the medium by
a randomized multiple of a slot time from 0 to (CW-1) to minimize the
chance of collision. During the random backoff interval, if the station
senses the medium becomes busy, it stops decrement the time counter and
does not reactivate the paused value until the channel is sensed idle again
for more than a DIFS. The size of CW is controlled by an exponential backoff mechanism. The value is set equal to a pre-specified minimum
contention window, CWmin, at the first transmission attempt. The CW is
doubled up to maximum contention window, CWmax, if the transmission
fails. Once the backoff timer expires, the transmitting station has two
options to initiate the transmission. It can either transmit the data packet
directly or transmit a short RTS (Request-To-Send) frame, followed by a
CTS (Clear-To-Send) frame from the receiving station. Furthermore, the
RTS-CTS frame contains the information of how long it takes to transmit
the data packet, which can help other stations to understand that they should
not attempt to transmit during such period. Once the data packet is
delivered to the destination station, an acknowledge (ACK) frame will be
sent back to confirm the transmission [18][19].
Point Coordination Function (PCF): The IEEE IEEE 802.11 MAC also
incorporates an optional access method called a PCF, which is only usable
on infrastructure network configurations. If the PCF is to be used, time is
divided into superframes where each superframe consists of a contention
period (CP) in which DCF is used and a contention-free period (CFP)
where PCF is used. This access method uses a point coordinator (PC),
which operates at the access point of the BSS, to determine which STA
29
currently has the right to transmit. The operation is done with the PC
performing the role of the polling master. The CFP is started when the PC
distributes information (using the ordinary DCF) within Beacon
management frames to gain control of the medium by setting the network
allocation vector (NAV) in STAs. During the CFP, the PC polls each
station in its polling list (the high priority stations), when they are clear to
access the medium. Stations that are polled must always respond to a poll.
A station being polled is allowed to transmit a data frame, and in case of
unsuccessful transmission the station may retransmit the frame after being
re-polled or during the next CP. If there are no pending transmissions, the
response from the STA is a null frame containing no payload. To ensure
that no DCF stations are able to interrupt this mode of operation, the
interframe space between PCF data frames (PIFS) is shorter than the DIFS.
To prevent starvation of stations that are not allowed to send during the
CFP, there must always be room for at least one maximum length frame to
be sent during the CP [19].
2.4.7 Management frames
The purpose of management frames is to establish initial communications
between stations and access points. Below we describe the management
frames relevant for our research:
Beacon frame: in an infrastructure network, an access point periodically
sends a beacon (according to the BeaconPeriod parameter) that provides
synchronization among stations utilizing the same BSS The beacon
includes a timestamp that all stations use to update what IEEE 802.11
defines as a timing synchronization function (TSF) timer. If the access point
supports the point coordination function, then it uses a beacon frame to
announce the beginning of a contention-free period. If the network is an
independent BSS, all stations periodically send beacons for synchronization
purposes. The beacons include information such as SSID, supported rates
and security parameters [19]. The beacon packets can be seen on Fig 3 and
fig 4 as are capture by Wireshark.
Fig 3 and fig 4 show the beacons transmitted by different Access Points
identified by their MAC addresses, with different sequence numbers and
flags.
From these flags, information such as supported rates, current channel,
30
Traffic Indication Map, country or vendor of the AP can be obtained.
Additional parameters such as operating mode of the AP, type of network,
CFP coordination capabilities, privacy or modulation used are also
depicted.
Fig. 3. Beacon tagged parameters
31
Fig. 4. Beacon fixed parameters
ATIM frame: a station with frames buffered for other stations sends an
announcement traffic indication message (ATIM) frame to each of these
stations during the ATIM window, which immediately follows a beacon
transmission. The station then transmits these frames to the applicable
recipients. The transmission of the ATIM frame alerts stations in sleep state
to stay awake long enough to receive their respective frames [19].
Authentication frame: a station sends an authentication frame to a station
or access point that it wants to authenticate with. The authentication
sequence consists of the transmission of one or more authentication frames,
depending on the type of authentication being implemented (open system or
shared key) [19].
De-authentication frame: a station sends a de-authentication frame to a
station or access point with which it wants to terminate secure
communications [19].
Probe request frame: a station sends a probe request frame to obtain
information from another station or access point. For example, a station
may send a probe request frame to determine whether a certain access point
is available [19].
32
Probe response frame: if a station or access point receives a probe request
frame, the station will respond to the sending station with a probe response
frame containing specific parameters about itself (such as parameter sets for
the frequency hopping and direct sequence PHYs)[19].
Association request frame: a station will send this frame to an access
point if it wants to associate with that access point. A station becomes
associated with an access point after the access point grants permission
[19].
Association response frame: after an access point receives an association
request frame, the access point will send an association response frame to
indicate whether or not it is accepting the association with the sending
station [19].
Re-association request frame: a station will send this frame to an access
point if it wants to re-associate with that access point. A re-association may
occur if a station moves out of range from one access point and within
range of another access point. The station will need to re-associate (not
merely associate) with the new access point so that the new access point
knows that it will need to negotiate the forwarding of data frames from the
old access point [19].
Re-association response frame: after an access point receives a reassociation request frame, the access point will send a re-association
response frame to indicate whether or not it is accepting the re-association
with the sending station [19].
2.4.1.6
Management procedures
Active scanning: consists on issuing probe request frames to which APs
will respond with probe response frames including information similar to
the information included in beacon frames. This way, the STA collects
information about candidate APs and chooses one of them. The selection of
which access point to use depends on several parameters such as quality of
signal, access network capabilities, user preferences and policy [19].
Passive scanning: the STA listens for beacons frames issued by the APs at
regular intervals [19].
33
Association: after a successful scanning in infrastructure mode, the STA
will try to associate with an AP. The STA will try to authenticate itself by
sending an authentication request to the AP, and the AP will respond with
an authentication response. If the authentication process is successful, the
STA sends an association request to the AP, and the AP responds with an
association response. If the STA meets the requirements to join the BSS that
the AP belongs to, the AP will include in the association response a status
“successful”. After the STA replies with an ACK, it will be associated with
the specific AP and it will be able to send traffic [19].
Re-association: when a station moves out of coverage of its associated AP,
it performs handoff procedures via scanning new APs and re-associating
with another AP. The STA initiates the same process as the one described in
the association procedure, but in this case the STA sends a re-association
request frame including the MAC address of the current AP, so that the new
AP can use it to communicate with the “old” AP for handoff procedures
[19].
Fig 5 shows the packets exchanged by an Access Point and a mobile device
during active scanning and association procedures.
Fig. 5. Scanning and association procedures
34
The active scanning is performed by a Samsung mobile device, which
sends out a Probe Request. The Access Point with the MAC address
Trapezen_71:13:84 responds (the first 3 bytes of the MAC address
correspond to the name of the device vendor) affirmatively with a Probe
Response, and the authentication process is started. After successful
authentication the mobile device requests association with the AP by
sending an Association Request, to which the AP responds positively with
an Association Response. Then, the AP asks the mobile device for the key
of the network, and once the mobile device provides this key (in this case is
automatically sent as it is stored in the mobile device memory), the data
traffic exchange starts.
2.4.2 Power saving mode (PSM)
PSM is the functionality that IEEE 802.11 networks utilize when there is no
active data traffic between the members of the network. The objective of
the IEEE 802.11 PSM is to let the wireless interface of a mobile host in the
active mode only for the time necessary to exchange data, and to turn it in
sleep mode whenever it becomes idle. As this master thesis deals with idle
modes in wireless networks, this is an important issue in our research.
Each mobile host within the hotspot informs the Access Point on whether it
utilizes the PSM or not by using the Power Management bits within the
Frame Control field of transmitted frame. Since the Access Point relays
every frame from/to any mobile host, it buffers the frames addressed to
mobile hosts using the Power-Saving Mode, and only transmit them at
designated times. PSM mobile hosts are synchronized with the Access
Point, and wake up to receive Beacons. The STAs that currently have
buffered MAC Service Data Units (MSDUs) within the AP are identified in
a traffic indication map (TIM), which is included as an element within all
beacons generated by the AP. The TIM indicates PSM mobile hosts having
at least one frame buffered at the Access Point, and a STA determines that
an MSDU is buffered for it by receiving and interpreting this TIM. This
information is coded in a partial virtual bitmap, whose bits are set by the
AP to select the STAs to which it has to forward pending buffered MSDUs.
If any STA in its BSS is in PS mode, the AP buffers all broadcast and
multicast MSDUs and deliver them to all STAs immediately following the
next Beacon frame containing a delivery TIM (DTIM) transmission. This
35
process is done using normal frame transmission rules, and before
transmitting any unicast frames to specific STAs [18].
If a STA is indicated in the TIM, it “downloads” the frames by sending a
special frame (PS Poll) to the Access Point by means of the standard DCF
procedure. Upon receiving a PS-Poll, the Access Point sends the first
DATA frame to the PSM mobile host, and receives the corresponding ACK
frame. If appropriate, the Access Point sets the More Data bit in the DATA
frame, to announce other frames to the same PSM mobile host. To
download the next frame, the mobile host sends another PS-Poll. If the TIM
indicating the buffered MSDU is sent during a contention-free period
(CFP), a CF-Pollable STA operating in the PS mode does not send a PSPoll frame, but remains active until the buffered MSDU is received (or the
CFP ends). A STA shall remain in its current PSM until it informs the AP
of a PSM change via a successful frame exchange. When, eventually, the
mobile host has downloaded all the buffered frames, it switches to the sleep
mode [18]. Fig 6 shows a typical PSM scenario.
.
Fig. 6. PSM scenario with DTIM at every 3 TIM [9]
36
CHAPTER
3
3
GSM/UMTS analysis
3.1 Brief description of USRP
The Universal Software Radio Peripheral (USRP) allows you to create
software radios using a computer with a USB 2.0, or Gigabit Ethernet port
regarding the version of the USRP. There are many daughter boards on
different radio frequency bands that can be used. For GSM and UMTS, we
used the DBSRX, which is a daughter board that can receive either GSM or
UMTS (800 – 2400 MHz). The prices of USRP’s vary regarding the
characteristics from $700 to $1700 dollars and daughter boards from $75 to
$475 dollars. The prices were taken based on [22]
The USRP works with GNU Radio where you can create your own
applications written in python programming language and the signal
processing is in C++.
For more detailed description of USRP’s characteristics and functions refer
to appendix 6.
One important limitation is that the USRP is capable of processing signals
up to 8 MHz wide. The uplink band for GSM is 25 MHz wide and for
UMTS is 60 MHz wide. Hence, if we wanted to collect data in a wider
band, we would have to pinpoint certain frequencies in order to listen to
either GSM or UMTS signals in a wider band.
The measurements for GSM and UMTS were performed at LTH in the E
building. It was observed that the activity located for the mobile phone was
within the first12 MHz of the uplink band for GSM (from 890 MHz to 902
MHz). While for UMTS, it was observed that the activity were located at
the end of the uplink band, so we scanned the last three channels (1967.4
MHz, 1972.4 MHz and 1977.4 MHz). These pre-measurements mentioned
before were performed with the spectrum analyzer.
37
It was scanned 12 MHz from 890 MHz to 902 MHz. It is important to
remember that the USRP can only scan 8 MHz at a time. Hence, it is
necessary to tune the center frequency as follows:


First step is 894 MHz (it will cover frequency band from 890 MHz
to 898 MHz).
Second step is 902 MHz (it will cover frequency band from 898
MHz to 906 MHz).
It was important to use FTT overlapping to reduce the non linearity
response of the digital down converter (frequency is not flat from –Fs/2 to
+Fs/2). So, it was chosen to use an overlap of 50%, this means that the step
size would be 4MHz. Thus, it was necessary to tune the center frequency as
follows:



First center frequency 892.7 MHz (frequency band from 888.7 MHz
to 896.7 MHz).
Second center frequency 896.7 MHz (frequency band from 892.7
MHz to 900.7 MHz).
Third center frequency 900.7 MHz (frequency band from 896.7
MHz to 904.7 MHz).
To do this, it was used the usrp_spectrum_sense.py (generally found in
gnuradio/gnuradio-examples/python/usrp/usrp_spectrum_sense.py). This
program is from the GNU radio software. In order to run it a proper
command would be ./usrp_spectrum_sense.py –f initial frequency final
frequency. The output of the command will be N samples point of the FFT
at different center frequencies without FFT overlapping. For more
information about the output of the command refer to appendix 6.
For the specific purposes, some lines were added to the script
usrp_spectrum_sense.py to support more than one USRP connected to one
computer. On the other hand, it was added some functions for FFT
overlapping, rearrange the output, and print it in the following order from
x[512] to x[1023] followed by x[1] to x[511] :
def adjust_data(m_data,fft_size):
It takes the output of the FFT
(m_data) and the size (fft_size) as arguments of the function. It returns the
output of the FFT from –Fs/2 to +Fs/2.
def overlapping(f_cent_1,f_resol,n_freq,f_step,f_size,t_data): It takes
38
the first center frequency (f_cent_1), frequency resolution (f_resol) that for
the measurements was 7812.5 MHz, the number of center frequencies
(n_freq) that will be used to scan the desire spectrum, the frequency step
(f_step) that was 4 MHz, the FFT size (f_size) which 1024 was chosen, and
the data arranged by adjust_data() (t_data) as arguments of the function. It
performs the FFT overlapping (50%) and returns the result.
def write_to_file(data,y): It writes the data in a text file so it can be
processed by Matlab.
To collect in a specific radio frequency band, type the following command
in the terminal:
./usrp_spectrum_sense.py -g xx --tune-delay xx --dwell-delay xx -F xx -d
xx xx.
-g xx: it is the gain (FPGA gain + RF amplifier gain).
-- tune-delay xx: When the USRP changes its center frequency, the USRP
has to wait until right number of samples arrive to our FFT engine and be
sure that it belongs to the desire center frequency.
--dwell-delay: it is the time that takes to collect N=1024 number of
samples with decimation rate = 8 is 128 us.
--F xx: it is the FFT size.
--d xx xx: it is the initial and final frequency.
Finally, in order to collect data for GSM, the following command was
executed in the terminal: ./usrp_spectrum_sense.py -g 40 --tune-delay 0.050
--dwell-delay 0.000130 -F 1024 -d 8 888.7M 904M.
In a similar way, it was executed the following command in the terminal for
UMTS: ./usrp_spectrum_sense.py -g 40 --tune-delay 0.050 --dwell-delay
0.000130 -F 1024 -d 8 1.9654G 1.9794G.
3.2
Active mode
It is important to check how often the mobile phone sends signals in the
uplink channel. In order to reduce user intervention, it was important to
study the idle mode state of the mobile phone, which is when the user
doesn’t interact with the mobile phone.
It was important to verify that the USRP was collecting correct data with all
39
the modifications done to usrp_spectrum_sense.py. The easiest way to
prove it was during the active mode state. It was performed an attempt call
for GSM and UMTS from one mobile phone to another and raw data was
collected with the USRP.
3.3
GSM and UMTS Attempt call
As it was mentioned before the measurement procedure for an attempt call
for either GSM or UMTS was performed as it can be seen in the following
picture.
Fig. 7. USRP, mobile phone and host
It is important to remind a couple of technical details for GSM and UMTS.
The uplink band goes from 890 to 915 MHZ for GSM. GSM uses slow
frequency hopping and in order to overcome this problem for collecting
data was measured a wider band range. For the measurements was taken
only the first 12 MHz as it was observed activity with the spectrum
analyzer within the first 12 MHz for that specific position in that specific
geographic area. GSM uses FDMA to operate in various frequencies
separated 200 KHz. The GSM spectrum within the first 12 MHz of the
uplink band can be seen in fig. 8 for an attempt call. There are many
40
frequencies in which the signal hop and each carrier frequency is separated
200 KHz each. This can be seen more clearly in fig. 9.
Fig. 8. GSM uplink spectrum
For UMTS, the uplink band is from 1920 to 1980 MHz. Regarding the
specific position and geographical area, it was observed that the activity for
UMTS was located in the last three channels of the band in the spectrum
analyzer (1967.4 MHz, 1972.4 MHz and 1977.4 MHz). Each channel is 5
MHz wide and as it uses spread spectrum, signal energy is distributed along
the frequency band.
In the fig. 10 can be seen that the signal activity was located on the carrier
frequency1972.4 MHz. In addition, the signal energy is distributed along 5
MHz (+2.5 MHz and -2.5MHz from the carrier frequency). Besides the
UMTS spectrum, it can be seen how the signal looks in the time domain.
Fig. 11 shows 9.996 seconds of an attempt call. It can be observed clearly
where it was signal activity as it produces a change in power levels.
41
Fig. 9. Frequency carrier separation of 200 KHz for GSM
Fig. 10. UMTS frequency spectrum (attempt call)
42
Fig. 11. UMTS time domain (attempt call)
3.4 IDLE mode
In order to reduce user intervention, it was required signal activity during
the idle mode. It is important to remember that the idle mode is when the
user is not using the mobile phone and there is no activity at all.
There are many tasks involved in the idle mode, furthermore, periodic
updates are important to keep the network updated; moreover, as battery
life is an important issue in most mobile devices, there has been a big effort
to enhanced battery life.
How often periodic signals are triggered is controlled by the network
operator. One main disadvantage for the purposes is the lack of regular
activity during the idle mode. The time of how often periodic signals are
triggered vary from 0 to 255 deci-hours (from six minutes to 25.5 hours).
Fig. 12 shows the signal activity in the idle mode state for GSM during
1858.321283 seconds (approximately 30 minutes), while fig. 13 shows the
same, but for UMTS during 1884.056545 seconds (approximately 31
43
minutes). Based on the fact that there is not continuously signal activity
during the idle mode state, we rather focused on WI-FI.
Fig. 12. GSM idle mode
Fig. 13. UMTS idle mode
44
CHAPTER
4
4
Implementation
4.1 USRP approach
Due to the nature of Wi-Fi, each channel is shared between upstream and
downstream. This turns into a challenge of how to separate between the
streams as Wi-Fi does not use Frequency Division Multiple Access
(FDMA).
Wireshark is a network analyzer that is based on layers 2 and 3. It means
that it captures packets and decodes them showing their MAC addresses
and power levels. More details about network analyzers will be discussed in
4.2.
The USRP captures incoming signals and collects their power levels. It
captures signals on layer 1 (pure physical layer).
In this section the objective is to distinguish between upstream and
downstream, and our approach will be to separate them through power
levels. It means that if it is placed the USRP close to a router, the USRP
will get higher power levels on the downstream than on the upstream.
4.1.1 Detecting IEEE 802.11 signals using USRPs
It is important to remember that Wi-Fi channels for IEEE 802.11 b and
IEEE 802.11 g/n are 22 MHz and 20 MHz respectively. Do not forget that
USRP can process signals only 8 MHz wide. There is a WI-FI network
installed in all corridors along the building. Our measurements were
performed at the basement in the E-building; furthermore, the router closest
to us was configured to transmit over channel 9.
In order to collect data with the USRP’s, it was followed the same method
as it was used either for GSM or UMTS. It was just tuned the receiver to
45
different center frequencies to cover the whole desire bandwidth. For the
same purpose, it was used the same code usrp_spectrum_sense.py with
proper parameters.
The command to be executed was the following:
./usrp_spectrum_sense.py -g 30 --tune-delay 0.010 --dwell-delay 0.000128
-F 1024 -d 8 2.439G 2.454G
If you wanted to collect data in a different channel you would just need to
change the last to parameters of the command to the correct initial
frequency and final frequency. For more details check appendix 6.
4.1.2 Scenario of measurements
The assumption was to place the USRP’s close to the router in order to get
higher power levels coming from the router than from the terminal.
Fig. 14 shows our measurement set up. The USRP1 was located right next
to the Wi-Fi access point, the USRP2 3.5 meters away from the USRP1.
The distance was limited by the length of the USB. Our hypothesis to get
higher power levels coming from the router than from the terminal only
applied to USRP1. Due that the USRP2 was 3.5 meters away, power levels
coming from the WI-FI access point were similar to the ones coming from
the terminal (making the distinction between streams challenging).
4.1.3 Procedure of measurements
In order to know which power levels were collected from the Wi-Fi access
point to both USRP’s, all incoming signals were measured to both USRP’s
to get a reference level of what we called noise (without triggering
upstream traffic from our terminals). Two reference levels for each USRP’s
were obtained.
For USRP1 two thresholds were defined: higher than -40 dBm and lower
than -60 dBm was captured most of the noise (downstream traffic), so the
power levels between – 40 dBm and -60 dBm was considered for upstream
traffic.
For USRP2 two thresholds were defined: higher than -45 dBm and lower
46
than -55 dBm was captured most of the noise (downstream traffic), so the
power levels between -45 dBm and -60 dBm was considered for upstream
traffic.
The measurement procedure was the following:




Stand up on different positions (POS 0, POS 1 and POS 2).
In each position, attempt a call and record data with both USRP’s.
Regarding the thresholds obtained in the pre-measurements, filter
only upstream traffic and plot it.
Plot in color red upstream traffic and in color blue downstream
traffic.
\
Fig. 14. A USRP Measurement set up
Fig. 15 shows the upstream traffic when the mobile phone was in POS 0
according to fig. 14 under USRP1. According to our thresholds for USRP1,
the power levels between -40 dBm and – 60 dBm was captured most of the
upstream traffic. As the USRP1 was closer to the WI-FI access point than
the mobile phone, the USRP1 captured higher power levels on downstream
traffic than upstream traffic from the mobile phone.
Fig. 16 shows less upstream traffic captured by USRP1. According on fig.
14, the mobile phone on POS 1 was further compared to POS 0 to USRP1.
That means that there might be upstream traffic which was weaker than
POS 0 and was not totally captured within our thresholds (between -40
47
dBm and -60 dBm).
Fig. 17 shows even less upstream traffic captured by USRP1. According on
fig. 14, the mobile phone was even further than POS1 to USRP1. Even
weaker signals could be captured on POS2 by the USRP1. Many incoming
signals detected by the USRP1 were not in our fixed thresholds.
The USRP2 was located according to fig. 14 approximately 3.5 meters
away from the Wi-Fi access point. In this case, the USRP2 was not
anymore receiving higher power levels from the Wi-Fi access point
compare to the upstream traffic. Regarding the position of the mobile
phone, upstream traffic could be less, similar or higher than the downstream
traffic. This is not good as we have a fix threshold (between -45 dBm and 55 dBm) where upstream traffic should have been.
Fig. 18 and fig. 19 shows the upstream captured by the USRP2 when the
mobile phone is on POS 0 and POS 2 according on fig. 14. Notice that
when the mobile phone was on POS 1, the USRP2 captured less upstream
traffic than POS 0. The reason was that the USRP2 as was not next to the
Wi-Fi access point, power levels captured by the USRP2 were not high
enough compared to power levels of upstream traffic on POS 1. This can be
seen on fig. 20 that there might be upstream traffic that was not within our
thresholds and it was considered as downstream traffic.
Fig. 15. USRP 1 pos 0, Wi-Fi traffic (Upstream in color red and downstream in color blue)
48
Fig. 16. USRP 1 pos 35, W-Fi traffic (Upstream in color red and downstream in color
blue)
Fig. 17. USRP 1 pos 45, Wi-Fi traffic (Upstream in color red and downstream in color
blue)
49
Fig. 18. USRP 2 pos 0, Wi-Fi traffic (Upstream in color red and downstream in color blue)
Fig. 19. USRP 2 pos 35, Wi-Fi traffic (Upstream in color red and downstream in color
blue)
50
Fig. 20. USRP 2 pos 45, Wi-Fi traffic (Upstream in color red and downstream in color
4.2 DD-WRT wireless broadband routers approach
In order to detect IEEE 802.11 signals with more accuracy an alternative
approach using broadband WLAN routers was utilized. This method let us
sense the 2.4 – 2.5 GHz spectrum to capture IEEE 802.11 packets for being
further analyzed with Wireshark and Matlab. A number of 3 routers are
used since this is the minimum number of nodes required for being able to
position a terminal.
The routers chosen for this purpose were Linksys WRT54GL (See appendix
C). This type of routers is Linux based and are compatible with DD-WRT
firmware (See appendix D). DD-WRT is a Linux based alternative
OpenSource firmware suitable for a great variety of WLAN routers and
embedded systems. The main emphasis lies on providing the easiest
possible handling while at the same time supporting a great number of
functionalities within the framework of the respective hardware platform
used. Compared to the software preinstalled on many WLAN routers, DDWRT allows a reliable operation with a clearly larger functionality that also
fulfills the demands of professional deployment. The main advantages of
the DD-WRT firmware are that it permits the installation of certain
programs in its hard disk, and that it provides a key functionality to the
routers, which is the capability of monitor the IEEE 802.11 network.
51
To be able to analyze the packets captured using the DD-WRT routers, the
packet network analyzers Wireshark and tcpdump are combined. With these
tools we can see the packets exchange between the members of a BSS
network, so that they can be further processed in Matlab.
4.2.1 Scenario of measurements
The measurements were performed at a typical open space office
environment. A map of the scenario is depicted on fig. 21:
Fig. 21. Map of the scenario
The scenario is a rectangular area of 10 x 25 meters that covers most of the
space occupied by the office. The grid is set to squares of 2.5 x 2.5 meters,
and every spot is numbered from 0 to 54. The origin of coordinates is the
spot number 0 (0, 0). Although every spot is separated 2.5 meters from the
ones surrounding it, we will just perform measurements in points separated
5 meters. This is suitable distance between two points so that the power
levels measured at these points are different enough to make a distinction
between the two spots regarding the power levels.
52
As it can be observed on the map, router 1 (R1) is placed on spot 20,
corresponding to coordinates (0 m, 10 m); router 2 (R2) is placed on spot
42, corresponding to coordinates (5 m, 20 m), router 3 (R3) is placed at
coordinates (8.95 m, 12.5 m). The height which the routers is placed at is
fixed to 1.05 meters. We would have liked to place the routers at a larger
height to avoid signal blocking in some parts of the scenario, but it was
physically impossible due to the structure of the measurements scenario.
The selection of the position of the routers is based on two constraints:
good coverage of the area and variability of power levels received at
different routers, which enhances indoor positioning algorithms
performance. The first premise is fulfilled since the routers give IEEE IEEE
802.11 network access at every spot on the office, and the second is also
fulfilled since the distance between routers is large enough to distinguish
power levels.
The positions at which the measurements took place were spots 18 (7.5 m,
7.5 m), 26 (2.5 m, 12.5 m), 28 (7.5 m, 12.5 m), 36 (2.5 m, 17.5 m), 38 (7.5
m, 17.5 m), and 45 (2.5 m, 22.5 m).
4.2.2 Detecting IEEE 802.11 signals using DD-WRT wireless
routers and packet network analyzers
As mentioned before, DD-WRT firmware gives LinkSys routers the
capability of have small programs installed on their hard disk, but also the
functionality of monitor every IEEE 802.11 signal in its reach. There are
three basic steps in order to detect and analyze IEEE 802.11 packets:
1) The command-line packet network analyzer tcpdump is installed in
each router (see Appendix E) and the wireless card each router is set
to work in “monitor” mode. This action creates the prism0 interface
in the router, which is able to detect every packet transmitted
through the IEEE 802.11 air interface.
2) Setting different filters on tcpdump allows discarding unnecessary
packets that do not belong to our research. As we are interested on
uplink data coming from the mobile phones, we set a specific filter
including the MAC addresses of the two mobile devices so that the
routers just capture the data that they send. Now we are ready to
capture packets.
53
3) The result of sensing the IEEE 802.11 spectrum with each router is
saved in a pcap file in the memory of the router. We send out each
file to a computer via LAN, where they are analyzed with another
packet network analyzer, Wireshark. With this program we are able
to see the data flow that we have captured, with detailed information
in every packet related to the IEEE 802.11 protocol. To be able to
easily to make calculations and graphs with these files, they are
convert to .txt files so that they can be processed with Matlab.
4.2.3 Limitations and assumptions
The routers R1, R2 and R3 form an infrastructure IEEE 802.11 network.
The router R1 is connected to the wired network, and the routers R2 and R3
are connected over Ethernet to R1. Thus, all routers work as AP.
When an AP is configured to be connected a specific channel, the prism0
interface that captures the IEEE 802.11 traffic can just detect the packets
sent over that channel. Since our purpose is to receive the same packets at
each router so that distinction made by power levels received from the same
packet at different positions can be done, each router has to be connected to
the same channel. This wireless network configuration is not suitable for
been deployed for operational purposes, i.e. public networks with thousands
of mobile terminals, since there would be channel overlapping leading to
interference issues and the performance would decrease significantly. To
our practical experience, the performance of the network was not affected
by this fact when 4 terminals were connected to it.
However, the routers can be switched to “repeater” mode, so that they are
simple sniffers that do not give Internet access, but can perfectly detect
IEEE 802.11 signals in the whole IEEE 802.11 b/g bandwidth. In this case,
an alternative infrastructure of APs operating in different channels (for
instance channels 1, 6 and 11), is assumed to be deployed to give Internet
access to users.
4.2.4 Procedure of measurements
Two smartphones (a Samsung and a HTC) were used to test the uplink
signals behavior when the mobile phone is connected to our Wi-Fi network,
but in idle or inactive state. As mentioned before, the spots where the
54
measurements took place were 18, 26, 28, 36, 38, and 45. We assumed that,
being the phone in Wi-Fi idle mode (corresponding with the Power Saving
Mode of IEEE 802.11), the most common places where a person has the
phone is in the pocket of a pants/trousers, and also holding it with the hand,
with an orientation parallel to the floor. From now on, we will refer to these
positions as “pocket” and “standard”. In order to test the influence on RSS
by the phone´s antenna, four different directions were added to these
positions. In both “pocket” and “standard” positions, directions “0
degrees”, “90 degrees”, “180 degrees” and “270 degrees” were tested to
involve the antenna radiation pattern factor to our measurements. In total, 8
measurements were undertaken for each spot in the map and for each
mobile phone. The 4 different directions are depicted on fig. 22, together
with the map of the scenario.
According to some previous experimental tests, the mobile devices
normally send out an average of 2 or 4 packets per minute when operating
in Wi-Fi Power Saving Mode, depending on the vendor. Counting the 6
different spots, 2 different positions and 4 different directions, the number
of measurements is 48 for each mobile phone. As the measurements
procedure requires active intervention by a person to recreate a real
scenario, and since this person has to hold one mobile phone with the hand,
having the other phone in the pocket at each corresponding spot, position
and direction, we assumed that 5 minutes were suitable to received enough
packets to make averaging over time, but also to not prolong considerably
the duration of the measurements, which was already 240 minutes (4 hours)
without counting time between measurements, etc.
In order to test the reproducibility and repeatability of the measurements
(important factor for the fingerprint technique), we performed the same
type of measurements a different day. An additional test was undertaken to
examine the possible difference in power levels between the uplink packets
sent by the phones when operating in PSM or when sending out data traffic.
For this purpose, the same measurement procedure (two smartphones,
different spots, positions and directions) was utilized to test the RSS at each
router when accessing a website (Facebook, for being one of the most
visited), watching a video, and making a Skype call. As in this mode the
mobile phone transmits larger amount of packets, shorter durations of the
measurements are needed.
55
Fig. 22. Orientations of the mobile phone used in the measurements
Routers R1, R2 and R3 operate in the same SSID and in channel 1, which
center frequency is 4.12 GHz. The beacon interval is 10 milliseconds and
DTIM is set to the default value, 3. As it will be demonstrated with the
experimental results, these values do not affect the periodicity of the uplink
signals.
56
CHAPTER
5
5
Results
5.1 USRP results
5.1.1 Distinction between upstream and downstream
Based on the measurements done with the USRP’s, it was observed that
trying to separate between upstream and downstream based on power levels
was quite challenging as IEEE 802.11 share the whole bandwidth for either
upstream or downstream not using FDMA . The physical layer is the same
for both streams and IEEE 802.11 does not make any difference using
different modulations, or techniques for a specific stream direction.
In addition, if it is desired to achieve the distinction based on power levels,
it could depend on how the Wi-Fi network is designed and where the
USRP’s are placed. Moreover, multipath effects could have the effect that
some signals will be affected by some fading dips for example and the
power levels do not fall within our thresholds. Furthermore, in some cases
some upstream that could not fall within our thresholds could be considered
as downstream.
One possible solution to the mentioned above, as USRP works with blocks
done in C++ and the linking between blocks is done with python, we could
develop a receiver for IEEE 802.11 and decode the information required for
the distinction.
The first two bytes of the MAC packet structure are composed by the frame
control. The frame control contains two flags which indicate if the traffic is
generated either by the access point or station.
We realized that decoding IEEE 802.11 packets could be done with simple
57
routers, hence, to develop an IEEE 802.11 receiver would be time
consuming, and it is out of the scope of our thesis work and timelines,
furthermore, about implementation cost would be reduced considerably
with just simple routers. So we decided to use routers instead of USRP.
5.1.2 Uniqueness of the signal
It is important to distinguish between upstream and downstream signals on
Wi-Fi traffic. Once, it can be done the distinction between upstream and
downstream, it is important to make unique each signal.
For this task, it was decided to use the MAC address as it is unique per
hardware. The MAC address does not change as the IP address. The MAC
address can be obtained by routers.
5.2 Results of measurements with IEEE 802.11 wireless
routers
5.2.1 Idle mode experimental behavior
One of the most important issues regarding the IEEE 802.11 uplink traffic
is the traffic pattern during PSM. As we are interested in locating a mobile
device by detecting the packets that it transmits, these packets have to be
transmitted within a time gap that has to be short enough to track the
mobility pattern of a person accurately. The figures presented below show
the PSM traffic pattern during the 5 minutes of a measurement for the two
different mobile devices.
From fig. 23 is observed that the uplink traffic pattern of the HTC is
characterized by the transmission of a packet every 15 seconds. The bit
PWR MGT is set to “1”, which means that the mobile device is operating in
Power Saving Mode. As the information transmitted is “Data”, we deduce
that the phone is responding to a DTIM beacon sent by the AP with a PSPoll frame containing data. We should also see the ACK packets sent by the
mobile device to the AP, but the MAC source address of these packets is not
specified so the Wireshark filter does not detect the ACK packets.
58
Fig. 23. HTC uplink traffic pattern 1
Fig. 24. HTC uplink traffic pattern 2
59
From fig. 24 is observed that the uplink traffic pattern of the HTC in this
case is different. Although it is not displayed, all the packets have the PWR
MGT bit set to “1”, so the mobile device is operating in Power Saving
Mode. In this case the traffic pattern of the HTS is characterized by the
following sequence:
Null function packet – around 7 seconds – Data packet – around 5 seconds
– Data packet – around 49 seconds – Null function packet (and the
sequence starts again).
When there are more than two data packets in a row, is due to failed
transmissions of a data packet, so it is retransmitted until the packet reaches
its destination (the retransmitted packets have the same sequence number).
From the figure it can be deduced that every 60 seconds (approximately),
the mobile device responds with a Null function packet to a previous DTIM
beacon sent by an AP, which means that the device does not have any data
packet to transmit. In between, and also with similar periodicity, two data
packets are transmitted by the mobile phone.
Fig. 25. Samsung uplink traffic pattern
60
The fig. 25 shows the IEEE 802.11 uplink traffic pattern of the Samsung
mobile phone. As the HTC, is operating in Power Saving Mode, but in this
case it transmits two consecutive Null function packets every 60 seconds
(approximately). The first packet informs the AP that the mobile device
does not have any information to transmit, and the second packet is also
Null function type, but it also informs the AP that it will go to sleep. After
60 seconds, the mobile device will repeat the same procedure. This
behavior is always the same, contrary to the HTC case.
As a conclusion we observe that there is a significant difference of behavior
between different vendors. The reason might be that each vendor follows
different power saving policies by setting different timers when operating in
IEEE 802.11 PSM. In any case both mobile devices send out enough
packets to be detected for the purpose of tracking mobility patterns.
Regarding the different traffic pattern that the HTC experiments, is worth to
say that the configuration of the network and the number of STAs
connected to it was exactly the same in all tests, so the reason for this
behavior might be the amount of battery that the mobile phone had at
different moments.
5.2.2 Variability analysis
Variability in power levels received at the different wireless routers is one
of the key factors of a successful indoor positioning technique. If there is a
remarkable difference in power levels received by each router at the
different spots, the performance of the positioning algorithm is enhanced
since the different positions computed depend on these power levels. In the
figures 26, 27, 28 and 29, it is demonstrated that the variability constraint is
fulfilled in our measurements:
The fig. 25, 26, 27, and 28 show the average power over 5 minutes received
at each router, at different spots, at different positions, at different
orientations and for both mobile phone models. It can be noticed that there
is a great variability of power levels depending on the router and the
orientation of the mobile phone. This fact reflects the influence of the
antenna radiation pattern on the measurements, and it adds more diversity
to the measurements. The RSS at each router also varies depending on the
distance to the different spots, according to a radio propagation scenario.
61
Fig. 26. Average power-Samsung-spot 26-standard position
TABLE 3. POWER STANDARD DEVIATION (dB)-SAMSUNG-SPOT 26-STANDARD
POSITION
Router 1
Router 2
Router 3
0 degrees
0,700
1,982
1,658
90 degrees
1,467
1,414
2,759
62
180 degrees
0
0,816
0,849
270 degrees
1,994
1,378
0,699
Fig. 27. Average power-Samsung-spot 28-pocket position
TABLE 4. POWER STANDARD DEVIATION (dB)-SAMSUNG-SPOT 28-POCKET
POSITION
Router 1
Router 2
Router 3
0 degrees
1,911
0,520
0,527
90 degrees
1,173
1,513
0,816
63
180 degrees
2,949
0,967
0,953
270 degrees
2,118
0,516
0,699
Fig. 28. Average power-HTC-spot 36-pocket position
TABLE 5. POWER STANDARD DEVIATION (dB)-HTC-SPOT 36-POCKET
POSITION
Router 1
Router 2
Router 3
0 degrees
1,053
0,977
1,643
90 degrees
0,848
0,455
1,759
64
180 degrees
1,761
0,936
1,712
270 degrees
1,671
1,329
1,667
Fig. 29. Average power-HTC-spot 38-standard position
TABLE 6. POWER STANDARD DEVIATION (dB)-HTC-SPOT 38-STANDARD
POSITION
Router 1
Router 2
Router 3
0 degrees
1,424
1,306
2,268
90 degrees
2,123
1,268
1,658
180 degrees
1,820
1,647
1,197
270 degrees
2,539
1,781
1,279
The tables 3, 4, 5, and 6 represent the corresponding standard deviation of
power levels, due to averaging over 5 minutes. It can be observed that this
deviation is in the range 1- 2 dB, which means that the time chosen for the
measurements is good enough to make an accurate average of power levels.
The received power levels at each router from each spot are averaged over
the 4 different directions, to give an overall idea of the variability of RSS
omitting the antenna factor. Figure 30, 31, 32 and 33 show this behavior.
65
Fig. 30. Average power-Samsung-all spots-standard position
Fig. 31. Average power-Samsung-all spots-pocket position
66
Fig. 32. Average power-HTC-all spots-standard position
In these graphs the great variability of power levels is confirmed. The RSSs
at each router differs considerably between them for both standard and
pocket positions.
There is also a slight variability between the power levels received when
the phone is in pocket or standard position. This is normal since they have
different heights.
Regarding the difference of power levels received from the two
Smartphones, the total average power received from the HTC is -60.09
dBm, while the average power received from the Samsung is -62.55 dBm.
As the IEEE 802.11 standard jus sets a maximum transmitted power of the
mobile devices, different vendors can vary the transmitted power value to
fit the features of the device according to the different power saving
policies that they follow. From our point of view, the vendor of the mobile
device shall not be an important factor to consider when it comes to
transmitted power levels.
67
Fig. 33. Average power-HTC-all spots-pocket position
The main conclusion obtained from the analysis of the variability of
received power levels is that these RSS are different enough for being used
within any indoor positioning technique.
5.2.3 Reproducibility analysis
Reproducibility is another important requirement for a successful indoor
positioning method. The received power levels at any time at the routers
from the same device must be similar so that there is no dependence of
when the measurements are performed. Also, there must be no dependence
of the type of traffic that the mobile device is transmitting, since we assume
(according to the information provided by the IEEE 802.11 standards) that
the power transmitted by a mobile phone when transmitting IEEE 802.11
packets does not depend of the type of traffic. Obviously, the mobile device
transmits more packets when accessing a website, downstreaming a video
or making a Skype call, but this fact shall not affect the transmitted power
in every packet.
68
In the figures depicted below, relevant comparisons between same
measurements performed at different days are presented. In addition, a
comparison between the power levels received in Power Saving Mode (idle
mode) and data traffic is also depicted.
In order to reduce time employed in the measurements, RSS at each router
with the phone in standard position at spots 26, 28, 36 and 38 were tested.
On fig. 34, fig. 35, fig. 36, and fig. 37 four graphs compare the same
measurements performed under the same conditions but at different days.
As it can be observed, the power levels patterns are pretty similar,
confirming the reproducibility of the measurements over time.
Fig. 34. Average power levels at different days-idle mode-router 2-Samsung-spot 36
69
Fig. 35. Average power levels at different days-idle mode-router 3-Samsung-spot 36
Fig. 36. Average power levels at different days-idle mode-router 1-HTCspot 28
70
Fig. 37. Average power levels at different days-idle mode-router 2-HTC-spot 28
Some examples of the comparison of RSS with different data traffic are
presented next:
Fig. 38. Average power levels at different days-idle vs traffic-router 1-Samsung-spot 38
71
Fig. 39. Average power levels at different days-idle vs traffic-router 2-Samsung-spot 26
5.2.4 Practical application using trilateration
The measurements and trilateration technique were applied in an office
environment. In a normal trilateration scenario, one device receives power
levels from three transmitters, and applying radio propagation models the
distances between each transmitter and the receiver are calculated. Using
the geometry of the three circles with radii equal to the distances calculated
previously, the position can be estimated. In our research we have
employed the most basic trilateration algorithm, given by the functions (4),
(5), and (6).
The coordinates of router i (i =1, 2, 3) are known (xi, yi, zi) and the
distances between each router and the corresponding spot (di) are computed
using a radio propagation model for indoor environments. Solving the
72
system of equation for each spot, the coordinates x, y, z (although z
coordinated is not relevant for our purpose and it will not be computed)
corresponding to the estimated position of the spot are obtained. In our
scenario, although the role of the participants is reverse (one single
transmitter, three receivers), the geometry constraints are the same as in a
normal trilateration scenario. The figure below shows a sketch of the
scenario
Fig. 40. Reverse trilateration scenario
The system of equations with numerical values of xi, yi, zi is:
…………………………….(7)
…………………………….(8)
……………………….(9)
The radio propagation model employed is the ITU site-general model for
path loss prediction in indoor environments [20]:
73
LTOTAL = 20log10f + Nlog10d + Lf (n) – 28................................................(10)
where f is the frequency in MHz, N is the power decay index, d is the
distance in meters (d > 1 meter), Lf(n) is the floor loss propagation factor
and n is the number of floors (n > 1). As in our case n=1, Lf (n) is equal to
0. Thus, the radio propagation model for 1 floor is as follows:
LTOTAL = 20log10f + Nlog10d – 28 ……………………………………….(11)
The path loss can be also expressed as the difference in dB between the
transmitted and the received power:
LTOTAL = PTX - PRX ……………………………………………………….(12)
By merging equations 11 and 12 we can express the received power PTX
with the following formula:
PRX = PTX - 20log10f - Nlog10d + 28 ………………………………………….(13)
For each measurement, we obtain three levels power levels at each router
from a distance d to the phone, which is located at a specific spot in the
map:
PRX1 = PTX - 20log10f - Nlog10d1 + 28………………………………………..(14)
PRX2 = PTX - 20log10f - Nlog10d2 + 28………………………………………..(15)
PRX3 = PTX - 20log10f - Nlog10d3 + 28………………………………………..(16)
As PTX and f are fixed values common to equations (14),(15) and (16), we
can create a system of 3 equations with 4 unknown factors by combining
the equations as follows:
PRX1 - PRX2 =Nlog10(d2/ d1)…………………………………………………….(17)
PRX2 - PRX3 =Nlog10 (d3/ d2)……………………………………………………(18)
PRX3 - PRX1 =Nlog10 (d1/ d3)…………………………………………………...(19)
From equations (4), (5) and (6) we know that the distance is a function of
the coordinates x, y and z. As we are not interested in the coordinate z, we
can write d1, d2 and d3 as a function of x and y, and include these
74
expressions in the equations (17), (18) and (19), creating the following
system:
……………………………….(20)
…………………………..(21)
…………………………...(22)
As a result, in equations (20), (21) and (22) we have a system of three
equations with three unknown factor which are x, y and N.
For spots 26, 28, 36 and 38 in the map we collected 40 vectors of received
powers at the routers (PRX1, PRX2, PRX3). These power levels are formed by 4
sectors that contain 10 values representing each of the 4 directions
measured at each spot (0, 90, 180 and 270) degrees). Spots 18 and 45 are
not considered since they are not fully covered by the routers.
The computation of x, y and N was performed using the function fsolve of
Matlab, that uses the algorithm trust region dogleg to solve non-linear
equations. The algorithm needs a starting point or a “guess” of the unknown
factors to solve the equations. In a real scenario, fingerprinting could be
used to give the trust region dogleg algorithm the starting point, comparing
the received vector of power levels with a list of fingerprints linked to each
spot. The fingerprint that matches better the received vector of power levels
would give the x and y coordinates to be used as the starting point for the
trust region dogleg algorithm. As we would have just a few values in our
vector of power levels (in our case, 3), fingerprinting would not be effective
by itself. Thus, the combination of these two techniques may result in a
significant enhancement of the positioning technique.
Figs. 41 and 42 show a comparison between the real positions (marked with
squares) and the estimated positions (marked with asterisks). Each color
represents one spot.
75
Fig. 41. Estimated positions vs real position of each spot (HTC)
Fig 42. Estimated positions vs real position of each spot (Samsung)
76
The average distance deviation is 2.2626 meters, which represents a great
accuracy for an indoor positioning system. However, we believe that the
accuracy would increase with a larger number of receivers covering the
entire scenario, which at the same time would enable to position the phone
at more spots.
77
CHAPTER
6
6
Conclusions and future work
The main goal during this thesis project was to answer the three questions
presented in the abstract, which were the base of this project work.
First question: How often the mobile device sends out signals during idle
mode state for the different radio sources?
The analysis showed that GSM and UMTS do not present regular and
continuously signal activity during the idle mode (from 6 minutes to 25.5
hours). On the other hand, IEEE 802.11 presents more regular signal
activity during the sleeping mode state (an average of three packets per
minute) so, IEEE 802.11 can be used for tracking mobile phones in indoor
environments.
In the future work, it would be interesting to figure out the way to increase
the regularity of signal activity during idle mode state by changing the
channel on the routers. There should be a balance between battery life and
regular signal activity.
Second question: Are these uplink signals feasible for a general indoor
positioning technique?
Through our analysis, we concluded that the data acquired using DD-WRT
routers can be used for any indoor positioning technique, since the
requirements of variability and reproducibility are fulfilled. In the practical
analysis of trilateration, positive results were obtained as a starting point for
a deeper research in this model. The combination of fingerprinting and
trilateration is an attractive approach, and thus further
In the future work, it would be necessary to test the combination
fingerprinting and trilateration as it was suggested in this master thesis.
Combining such techniques would be a powerful and attractive approach
for indoor positioning. Furthermore, it would be definitely useful to
78
increase the number of routers/receivers and to simulate more advanced
algorithms to increase precision of the positioning technique. In addition,
radio environment analyzers like the Ekahau HeatMapper would give a
better estimation of the coverage of the receivers so that they can be placed
in an effective way.
Third question: Is there a way to find uniqueness between terminals?
We concluded that using the MAC address on IEEE 802.11 networks is
feasible to have uniqueness between different users.
To sum it up: IEEE 802.11 uplink signals can be used for indoor tracking.
We have been able to separate upstream and downstream traffic with simple
routers and DD-WRT software installed on them. Furthermore, there is no
need to have a software program installed on the mobile phone. We have
reduced user intervention. In addition, trilateration has shown fairly positive
results.
On the other hand, routers are a lot cheaper than USRP devices. When it
comes to implementation cost, it will require less investment on the
hardware.
One main disadvantage is that it is required that the Wi-Fi transceiver is
turned on. One solution might be to provide free internet in the desired area
to cover.
Another disadvantage is that in some mobile phones the default Wi-Fi sleep
policy configuration is set to last 15 min or until the screen is off. It means
that after 15 min or when the screen is off, the Wi-Fi transceiver will be
switched off. On the other hand, there are some other mobile phones where
the default Wi-Fi sleep policy is set as “never”, which means that Wi-Fi
will be operating all the time since it is turned on by the user until it is
turned off manually by the user.
79
References
[1] http://loclizard.com/products/loclizard/
[2] http://www.qubulus.com.
[3] A.K.M. Mahtab Hossain, Hien Nguyen Van, Yunye Jin, Wee-Seng Soh Indoor Localization Using Multiple Wireless Technologies. IEEE Press,
2007.
[4] Eddie C.L. Chan, George Baciu, S.C. Mak - Wireless Tracking Analysis
in Location Fingerprinting. IEEE Press, 2008.
[5] Rong-Hong Jan; Yung Rong Lee; , "An indoor geolocation system for
wireless LANs," Parallel Processing Workshops, 2003. Proceedings. 2003
International Conference on , vol., no., pp. 29- 34, 6-9 Oct. 2003
[6] http://en.wikipedia.org/wiki/Trilateration
[7] Kundu Sudakshina – Analog and Digital Communications. Pearson,
2010.
[8] Jorg Eberspacher, Jörg Eberspächer, Christian Bettstetter, Hans-Joerg
Vögel, Christian Hartmann, Hans-Joerg Vgel, Jö Rg Eberspä Cher - GSM Architecture, Protocols and Services. Wiley, 2009.
[9] Andreas F. Molisch - Wireless Communications. Wiley/IEEE Press,
2005.
[10] Thomas Toftegaard Nielsen, Jeroen Wigard-Performance
enhancements in a frequency hopping GSM network. Public Academic
Publishers, 2000.
[11] 3GPP TS 25.304 V10.0.0 User Equipment (UE) procedures in idle
mode and procedures for cell reselection in connected mode. 3GPP, 2011.
[12] 3GPP TS 43.022 V10.0.0 Functions related to Mobile Station in idle
mode and group receive mode. 3GPP, 2011.
[13] Dahlman, E.; Beming, P.; Knutsson, J.; Ovesjo, F.; Persson, M.;
80
Roobol, C.; , "WCDMA-the radio interface for future mobile multimedia
communications," Vehicular Technology, IEEE Transactions on , vol.47,
no.4,pp.1105-1118,Nov1998
[14] Wikipedia IEEE IEEE 802.11g-2003
http://en.wikipedia.org/wiki/IEEE_IEEE 802.11g-2003.
[15] Branimir Boskovic, Milan Markovic On Spread Spectrum Modulation
Techniques Applied in IEEE IEEE 802.11 Wireless LAN Standard. IEEE
Press, 2000.
[16] Moustaffa A. Youssef, Raymond E. Miller – Analyzing the Point
Coordination Function of the IEEE IEEE 802.11 WLAN Protocol using a
Systems of Communicating Machines Specifications. University of
Maryland, 2002.
[17] IEEE Standard IEEE 802.11g Information technology —
Telecommunications and information exchange between systems — Local
and metropolitan area networks — Specific requirements — Part 11:
Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY)
specifications. IEEE Press 2003.
[18] Shao-Cheng Wang, Yi-Ming Chen, Tsern-Huei Lee, Ahmed Helmy –
Performance Evaluations for Hybrid IEEE IEEE 802.11b and IEEE
802.11g
Wireless
Networks.
Performance,
Computing, and Communications Conference, IPCCC 2005.
[19] IEEE Standard IEEE 802.11 Information technologyTelecommunications and information exchange between systems- Local and
metropolitan area networks- Specific requirements- Part 11 : Wireless LAN
Medium Access Control (MAC) and Physical Layer (PHY) specification1997.
[20] J. Seybold, Introduction to RF Propagation, Wiley Interscience , 2005.
[21] www.ettus.com
81
List of Figures
Fig. 1. Components of the UMTS system ................................................... 22
Fig. 2. Uplink frame structure ..................................................................... 24
Fig. 3. Beacon tagged parameters ............................................................... 31
Fig. 4. Beacon fixed parameters.................................................................. 32
Fig. 5. Scanning and association procedures .............................................. 34
Fig. 6. PSM scenario with DTIM at every 3 TIM ...................................... 36
Fig. 7. USRP, mobile phone and host ......................................................... 40
Fig. 8. GSM uplink spectrum ...................................................................... 41
Fig. 9. Frequency carrier separation of 200 KHz for GSM ........................ 42
Fig. 10. UMTS frequency spectrum (attempt call) ..................................... 42
Fig. 11. UMTS time domain (attempt call) ................................................. 43
Fig. 12. GSM idle mode .............................................................................. 44
Fig. 13. UMTS idle mode ........................................................................... 44
Fig. 14. A USRP Measurement set up ........................................................ 47
Fig. 15. USRP 1 pos 0, Wi-Fi traffic (Upstream in color red and
downstream in color blue)........................................................................... 48
Fig. 16. USRP 1 pos 35, W-Fi traffic (Upstream in color red and
downstream in color blue)........................................................................... 49
Fig. 17. USRP 1 pos 45, Wi-Fi traffic (Upstream in color red and
downstream in color blue)........................................................................... 49
Fig. 18. USRP 2 pos 0, Wi-Fi traffic (Upstream in color red and
downstream in color blue)........................................................................... 50
Fig. 19. USRP 2 pos 35, Wi-Fi traffic (Upstream in color red and
downstream in color blue)........................................................................... 50
Fig. 20. USRP 2 pos 45, Wi-Fi traffic (Upstream in color red and
downstream in color .................................................................................... 51
Fig. 21. Map of the scenario ....................................................................... 52
Fig. 22. Orientations of the mobile phone used in the measurements ........ 56
Fig. 23. HTC uplink traffic pattern 1 .......................................................... 59
Fig. 24. HTC uplink traffic pattern 2 .......................................................... 59
Fig. 25. Samsung uplink traffic pattern....................................................... 60
Fig. 26. Average power-Samsung-spot 26-standard position ..................... 62
Fig. 27. Average power-Samsung-spot 28-pocket position ........................ 63
Fig. 28. Average power-HTC-spot 36-pocket position............................... 64
Fig. 29. Average power-HTC-spot 38-standard position ............................ 65
Fig. 30. Average power-Samsung-all spots-standard position ................... 66
82
Fig. 31. Average power-Samsung-all spots-pocket position ...................... 66
Fig. 32. Average power-HTC-all spots-standard position .......................... 67
Fig. 33. Average power-HTC-all spots-pocket position ............................. 68
Fig. 34. Average power levels at different days-idle mode-router 2Samsung-spot 36 ......................................................................................... 69
Fig. 35. Average power levels at different days-idle mode-router 3Samsung-spot 36 ......................................................................................... 70
Fig. 36. Average power levels at different days-idle mode-router 1HTCspot 28 ................................................................................................. 70
Fig. 37. Average power levels at different days-idle mode-router 2-HTCspot 28 ......................................................................................................... 71
Fig. 38. Average power levels at different days-idle vs traffic-router 1Samsung-spot 38 ......................................................................................... 71
Fig. 39. Average power levels at different days-idle vs traffic-router 2Samsung-spot 26 ......................................................................................... 72
Fig. 40. Reverse trilateration scenario ........................................................ 73
Fig. 41. Estimated positions vs real position of each spot (HTC) .............. 76
Fig. 42. Estimated positions vs real position of each spot (Samsung) ........ 76
83
List of Tables
TABLE 1. LIST OF IEEE 802.11 B/G CHANNELS ............................... 25
TABLE 2. DATA RATES IN IEEE 802.11g ............................................. 28
TABLE 3. POWER STANDARD DEVIATION (dB)-SAMSUNG-SPOT
26-STANDARD POSITION ...................................................................... 62
TABLE 4. POWER STANDARD DEVIATION (dB)-SAMSUNG-SPOT
28-POCKET POSITION............................................................................. 63
TABLE 5. POWER STANDARD DEVIATION (dB)-HTC-SPOT 36POCKET POSITION .................................................................................. 64
TABLE 6. POWER STANDARD DEVIATION (dB)-HTC-SPOT 38STANDARD POSITION ............................................................................ 65
84
Appendix
1
A.1 Python code
It was modified and added some new code to the script developed by GNU
radio. The script is generally found in gnuradio/gnuradioexamples/python/usrp/usrp_spectrum_sense.py.
It will be just shown the lines that were modified and the new methods and
code developed.
# build graph
if self.opc==1:
self.u =
usrp.source_c(which=0,fusb_block_size=options.fusb_block_size,fusb_nbl
ocks=options.fusb_nblocks)
if self.opc==2:
self.u =
usrp.source_c(which=1,fusb_block_size=options.fusb_block_size,fusb_nbl
ocks=options.fusb_nblocks)
if self.opc==3:
self.u =
usrp.source_c(which=2,fusb_block_size=options.fusb_block_size,fusb_nbl
ocks=options.fusb_nblocks)
# Set the freq_step to 50% of the actual data throughput.
# This allows us to discard the bins on both ends of the spectrum.
self.freq_step = 0.5 * usrp_rate
# add the following new methods
def adjust_data(m_data,fft_size): # adjust the fft from -Fs/2 to Fs/2
var1=m_data[0:1]
var2=m_data[1:fft_size/2]
85
var3=m_data[fft_size/2:]
var3.extend(var1)
var3.extend(var2)
return var3
def overlapping(f_cent_1,f_resol,n_freq,f_step,f_size,t_data):
#
Ovelapping to eliminate zeros in the fft
f_cent1_max=f_cent_1+f_resol*(f_size/2-1)
f_cent_2=f_cent_1+f_step
f_cent2_min=f_cent_2-f_resol*(f_size/2)
num_sample_up=int((f_cent1_max-f_cent2_min)/f_resol)
#num_sample_low=(f_cent2_min-f_cent_1)/f_resol
j=1
var=t_data[0:f_size-1-num_sample_up]
while j<n_freq:
g1=array(t_data[(f_size*j-1)-num_sample_up:f_size*j])
g2=array(t_data[f_size*j:f_size*j+num_sample_up+1])
ii=0
for x in g2:
if g1[ii]<=x:
g1[ii]=x
ii=ii+1
else:
ii=ii+1
#g_aver=list((g1+g2)/2)
#g_aver=list(g1+g2)
#var.extend(g_aver)
var.extend(g1)
var_h=t_data[f_size*j+num_sample_up+1:(f_size*(j+1)-1)num_sample_up]
var.extend(var_h)
if j==n_freq-1:
aux=t_data[f_size*n_freq-1-num_sample_up:]
var.extend(aux)
j=j+1
else:
j=j+1
return var
def write_to_file(data,y): # write to text file
86
if y==1:
f=open('/home/rick/collected_data/umts/umts_idle.dat','a')
if y==2:
f=open('/home/rick/collected_data/Wi-Fi/test/usrp2_45.dat','a')
if y==3:
f=open('/home/rick/collected_data/Wi-Fi/test/up_pos1_usrp3.dat','a')
k=str(data)
k=k.replace("["," ")
k=k.replace("]"," ")
f.write(k)
#f.write("\n")
f.close()
# The part below you can just replace as it contains several changes and
new code
def main_loop(tb):
while 1:
# Get the n,ext message sent from the C++ code (blocking call).
# It contains the center frequency and the mag squared of the fft
i=1
temp=[]
temp1=[]
temp2=[]
usrp=8e6
##### set parameters
freq_min=1.9654e9
freq_max=1.9794e9
fft_size=1024
freq_step=4e6
usrp2=False
usrp3=False
num_freq = math.ceil((freq_max - freq_min) / freq_step)
num_freq = int(num_freq)
#print num_freq
freq_cent_min=freq_min + freq_step/2
87
freq_resol=usrp/fft_size
while i<=num_freq:
m = parse_msg(tb.msgq.delete_head())
#print m.center_freq
if usrp2:
m1 = parse_msg(tb1.msgq.delete_head())
if usrp3:
m2 = parse_msg(tb2.msgq.delete_head())
#print m.data
#print "-----------------------------------------------------------"
#print m1.data
m = list(m.data)
a_data=adjust_data(m,fft_size)
temp.extend(a_data)
if usrp2:
m1 = list(m1.data)
a_data1=adjust_data(m1,fft_size)
temp1.extend(a_data1)
if usrp3:
m2 = list(m2.data)
a_data2=adjust_data(m2,fft_size)
temp2.extend(a_data2)
if i==num_freq:
over_data=overlapping(freq_cent_min,freq_resol,num_freq,freq_step,fft_si
ze,temp)
write_to_file(over_data,1)
temp=[]
if usrp2:
over_data1=overlapping(freq_cent_min,freq_resol,num_freq,freq_step,fft_s
ize,temp1)
write_to_file(over_data1,2)
temp1=[]
if usrp3:
over_data2=overlapping(freq_cent_min,freq_resol,num_freq,freq_step,fft_s
ize,temp2)
write_to_file(over_data1,3)
88
temp2=[]
i=1
else:
i=i+1
#Print center freq so we know that something is happening...
#print m.center_freq
# FIXME do something useful with the data...
# m.data are the mag_squared of the fft output (they are in the
# standard order. I.e., bin 0 == DC.)
# You'll probably want to do the equivalent of "fftshift" on them
# m.raw_data is a string that contains the binary floats.
# You could write this as binary to a file.
if __name__ == '__main__':
now = datetime.datetime.now()
print str(now)
tb = my_top_block(1)
#tb1 = my_top_block(2)
#tb2 = my_top_block(3)
try:
tb.start()
# start executing flow graph in another thread...
#tb1.start()
#tb2.start()
main_loop(tb)
except KeyboardInterrupt:
now = datetime.datetime.now()
print str(now)
pass
89
Appendix
2
A.1 Parser code
The following code was developed in order to parse text files with data
collected by the routers and get the information required by us.
import javax.swing.JFileChooser;
import javax.swing.JOptionPane;
import java.io.*;
import java.util.StringTokenizer;
import java.util.regex.Matcher;
import java.util.regex.Pattern;
class LogParser {
public static void main (String[] args) throws IOException{
int status;
File file;
String fileName;
String str;
//Open the FileChooser
JFileChooser chooser = new
JFileChooser("/home/rick/collected_data/Wi-Fi_traffic/malformed");
while(true){
status = chooser.showOpenDialog(null);
if(status ==
JFileChooser.APPROVE_OPTION){
file =
chooser.getSelectedFile();
break;
}else{
JOptionPane.showMessageDialog(null, "Select a file");
90
}
}
//
FileReader fileReader = new FileReader(file);
BufferedReader bufReader = new
BufferedReader(fileReader);
// Get output fileName
fileName = JOptionPane.showInputDialog(null,
"Enter output file name for htc");
Matcher matcher;
Pattern pattern = Pattern.compile("Malformed",
Pattern.CASE_INSENSITIVE);
Matcher matcherTime;
Pattern patternTime =Pattern.compile("Power",
Pattern.CASE_INSENSITIVE);
Matcher matcherPower;
Pattern patternPower =Pattern.compile("-[09][0-9]", Pattern.CASE_INSENSITIVE);
//
File outFile = new
File("/home/rick/collected_data/Wi-Fi_traffic/corregidos/"+fileName);
FileOutputStream outFileStream = new
FileOutputStream(outFile);
PrintWriter outStream = new
PrintWriter(outFileStream);
while(true){
String line = bufReader.readLine();
if(line == null) {
outStream.close();
break;
}
91
//
matcher = pattern.matcher(line);
matcherTime =
patternTime.matcher(line);
matcherPower = patternPower.matcher(line);
boolean booleanMalformed =
matcher.find();
boolean booleanTime =
matcherTime.find();
boolean booleanPower =
matcherPower.find();
if(false==booleanTime){
if(false==booleanMalformed){
System.out.println(line);
StringTokenizer strTokenizer = new StringTokenizer(line);
int j=2;
while(strTokenizer.hasMoreTokens()){
if(j==1){
outStream.print(strTokenizer.nextToken());
outStream.print(" ");
j++;
}
if(j==2){
if(true==booleanPower){
str = line.substring(matcherPower.start(),
matcherPower.start()+3);
92
outStream.print(str);
outStream.print("\n");
j++;
}
}
strTokenizer.nextToken();
j++;
}
} else{
System.out.println(booleanMalformed);
}
}
}
}
}
93
Appendix
3
A.1 LYNSYS WRT54GL specifications
Model: WRT54GL.
Platform: Broadcom MIPS
CPU: Broadcom BCM5452 at 200 MHz (130nm construction)
Standards: IEEE 802 .3, IEEE 802 .3u, IEEE 802 .11g, IEEE 802 .11b.
Channels: 11 Channels (US, Canada), 13 Channels (Europe, Japan).
Ports: Internet: One 10/100 RJ-45 Port ; LAN: Four 10/100 RJ-45
Switched Ports, One Power Port.
Flash: 4 MB NAND, single chip.
System Memory: 16 MB 16-bit DDR SDRAM.
USB: None.
Wireless Radio: Broadcom BCM43xx IEEE 802.11b/g.
Antenna: Dual folding, removable, rotating antennas.
Button: Reset, SecureEasySetup.
Cabling Type: CAT5.
LEDs: Power, DMZ, WLAN, LAN (1-4), Internet, SecureEasySetup
RF Power Output: 18 dBm.
UPnP able/cert: Able.
94
Security Features: Stateful Packet Inspection (SPI). Firewall, Internet
Policy.
Wireless Security: Wi-Fi Protected Access™2 (WPA2), WEP, Wireless
MAC Filtering.
Dimensions: 186 x 48 x 154 mm.
Weight: 391 g.
Power: External, 12V DC, 0 .5A.
Certifications: FCC, IC-03, CE, Wi-Fi (802 .11b, 802 .11g), WPA2,
WMM.
Operating Temperature: 0 to 40ºC.
Storage Temperature: -20 to 70ºC.
Operating Humidity: 10 to 85%, Noncondensing.
Storage Humidity: 5 to 90%, Noncondensing.
Information acquired at
http://downloads.linksysbycisco.com/downloads/userguide/1224639045490
/WRT54GL-EU_11_UG_A-WEB.pdf.
95
Appendix
4
A.1 Getting started with dd-wrt
Requirements




A computer (Windows, Linux, Mac, whatever).
A broadband internet connection (DSL, Cable, or similar).
A Linksys WRT54G/GL/GS router or other supported router.
The DD-WRT firmware image from The DD-WRT Project.
Information acquired at http://www.ddwrt.com/wiki/index.php/What_is_DD-WRT%3F.
Installation process from a stock Linksys firmware
1. READ the peacock announcement linked above.
2. Do a hard reset.
You can HARD RESET by holding down the reset button on the back of
the router for 30 seconds, then pulling the power cord for 30 seconds while
STILL holding the reset button, and then plugging in the power cord for a
final 30 seconds while STILL holding the reset button. You will hold the
reset button in for 90 seconds without releasing it. Then release the reset
button and wait for the router to finish doing whatever it's going to do.
Usually the WLAN light will come on close to last in the boot sequence.
Sometimes, however, the POWER light will keep flashing for a good while.
Either way, once you're sure the router has done its thing, power cycle the
router, by unplugging and re-plugging the power connector in the back of
the router. There's no need to wait between unplugging and replugging.
3.
Download the NEWD_MINI build generic 12548 Eko Build
from
here: ftp://dd-wrt.com/others/eko/V24_TNG/svn12548/ddwrt.v24-12548_NEWD_mini.bin
96
4.
You should check the MD5 HASH of the firmware after
downloading it if it is available for your build. See Hashes &
Checksums. Turn off and disable your firewall, turn off and disable
your antivirus, and sign into your linksys router with your web browser.
5.
Use the firmware upgrade web interface to update your router
with dd-wrt. DO NOT close your browser; DO NOT interrupt the
process, be EXTREMELY PATIENT, even after the firmware is
already supposedly upgraded. Wait around for a while, and make sure
it settles down and is definitely finished doing whatever it's going to do.
The router needs time to rebuild the NVRAM after it has been flashed,
and if you interrupt this you will regret it!
6.
When three or so minutes pass and the WLAN light comes
back on, try and open the dd-wrt page at 192.168.1.1. If it opens, the
process is complete. You should be asked to reset the password. Type
username and passphrase in carefully.
7.
Decide if you would like to keep the MINI version, or change
to MICRO, STANDARD, VOIP, or VPN versions. DO NOT try to
load a MEGA build on this router (see above). If you are keeping
MINI, you are done, otherwise, continue
Read about the different versions' features here: What is DD-WRT?#File
Versions. If you won't be needing the features in the larger versions such as
standard, you may be able to increase the responsiveness of your router by
getting the smallest version that includes the features you need. Also, you
can always update to a larger version later if down the line you need the
extra features.
8. Power cycle the router.
9. Do a hard reset.
10. Install the version of dd-wrt you want.
11. Wait again for the process to complete and the lights to return to
normal, and the webgui to be accessible at 192.168.1.1 showing the
updated firmware. Give it at least 3 to five minutes.
97
12. Power cycle the router.
13. Do a hard reset
Other install notes
If you are upgrading from the web interface, you should use the GENERIC
versions. For updating by webgui, EKO build 12548 Newd_Mini.bin is the
recommended build for this router. It works well. You can also upgrade to
12548 Newd_Std.bin AFTER you have put on the mini version. Here is a
link to the mini version download:
ftp://dd-wrt.com/others/eko/V24_TNG/svn12548/dd-wrt.v2412548_NEWD_mini.bin.
Information acquired at http://www.ddwrt.com/wiki/index.php/Linksys_WRT54GL
For further information visit the official DD-WRT website http://www.ddwrt.com/site/index.
98
Appendix
5
A.1 Code to install tcpdump on DD-WRT routers
#!/usr/local/bin/expect
spawn telnet 192.168.1.1 % The IP address here is the router IP ad
dress.
In our case we have 192.168.1.1, 192.168.1.2 and 192.168.1.3
expect "login:"
send "root\n"
expect "Password:"
send "admin\n"
expect "$>"
send "wl monitor 1\r"
send "mount --bind /tmp/smbshare /jffs\r"
send "nvram set sys_enable_jffs2=1 \r"
send "mkdir -p /tmp/smbshare/tmp/ipkg\r"
send "cd /tmp/smbshare/tmp/ipkg\r"
send "wget http://downloads.openwrt.org/whiterussian/packages/libpcap_0.9.41_mipsel.ipk\r"
send "ipkg -d smbfs install libpcap_0.9.4-1_mipsel.ipk\r"
send "wget http://downloads.openwrt.org/whiterussian/packages/tcpdump_3.9.41_mipsel.ipk\r"
send "ipkg -d smbfs install tcpdump_3.9.4-1_mipsel.ipk\r"
99
Appendix
6
A.1 Getting started with the USRP
This is a short guide of how to install GNU radio.
1) Build the USRP with the daughter boards you require for your
application
2) Install Ubuntu on your computer and download updates
3) Open the terminal and copy the following (for ubuntu 10.04):
sudo apt-get -y install libfontconfig1-dev libxrender-dev libpulse-dev
swig g++ automake autoconf libtool python-dev libfftw3-dev \
libcppunit-dev libboost-all-dev libusb-dev fort77 sdcc sdcc-libraries \
libsdl1.2-dev python-wxgtk2.8 git-core guile-1.8-dev \
libqt4-dev python-numpy ccache python-opengl libgsl0-dev \
python-cheetah python-lxml doxygen qt4-dev-tools \
libqwt5-qt4-dev libqwtplot3d-qt4-dev pyqt4-dev-tools python-qwt5qt4
4) Install GNU radio from git
git clone http://gnuradio.org/git/gnuradio.git
cd gnuradio
./bootstrap
./configure
make
100
5) if you have USRP1
sudo addgroup usrp
sudo usermod -G usrp -a <YOUR_USERNAME>
echo 'ACTION=="add", BUS=="usb", SYSFS{idVendor}=="fffe",
SYSFS{idProduct}=="0002", GROUP:="usrp", MODE:="0660"' >
tmpfile
sudo chown root.root tmpfile
sudo mv tmpfile /etc/udev/rules.d/10-usrp.rules
6) you can check if the USRP is connected with the following
command
ls -lR /dev/bus/usb | grep usrp
The result should appear as:
crw-rw---- 1 root usrp 189, 514 Mar 24 09:46 003
7) Once you have verified that the USRP is connected to the computer,
Now, it is time to check if the USRP works, Run the following
scripts:
Python interface:
cd gnuradio-examples/python/usrp
./usrp_benchmark_usb.py
C++ interface:
cd usrp/host/apps
./test_usrp_standard_tx
101
./test_usrp_standard_rx
If both have work, you are now ready to use GNU radio and USRP.
THE USRP
The Universal Software Radio Peripheral (USRP) connects to the computer
through USB. It is used for making software radios. Regarding your
application, you just need the right daughter board that should be plugged
into the either RX or TX slots on the USRP.
The USRP runs at 64 MHz, however, the computer cannot receive this
information fast enough. Using decimation is a way in which it can be
reduced the number of samples retrieve from the USRP. According to [3] it
can retrieve 8 MHz from the USRP due that the USB limitations.
There are new versions of the USRP, which are faster and can retrieve 16
MHz from the USRP.
If you want to sense the spectrum you can run the following command:
usrp_spectrum_sense.py (generally found in gnuradio/gnuradioexamples/python/usrp/usrp_spectrum_sense.py). The output of running the
command is explained below:
Assume you have N=1024 (I and Q) samples gathered using a tuner center
at 896.7 MHz. The sampling frequency is 8 MHz. Doing 1024 points
complex FTT means [21]:



Frequency resolution is: 8MHz /1024 = 7.8125 KHz.
The output of the FFT x[0] represents the spectrum at 896.7 MHz.
The output of the FFT x[1] to x[511] represents the frequencies
from 896.7078125 to 900.6921875 MHz.
The output of the FFT x[512] to x[1023] represents the frequencies from
892.7078125 MHz to 896.6921875 MHz
102
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertising