aspiringtokΤεχνίτη Νοημοσύνη και Ρομποτική

15 Οκτ 2013 (πριν από 3 χρόνια και 2 μήνες)

99 εμφανίσεις

M.Sc. Thesis




Assoc. Prof. Moustafa Youssef

Dr. Khaled A. Harras

Prof. Tatsuo Nakajima


Many indoor location determination technologies have been proposed over the years, including:
infrared, ultrasonic, and radio frequency (RF). All these technologies provide varying levels of accuracy
supporting different application needs. However, thes
e technologies require the deployment of special
hardware and/or special calibration of the area of interest. In this thesis, we address the problem of
realizing a ubiquitous indoor positioning system (IPS). Similar to the outdoor GPS, IPS is envisioned to

deployed on a large scale worldwide, with minimum overhead, to work with heterogeneous devices.
Such a system will enable a wide set of applications including direction finding between indoor
locations, enhancing first responders' safety, etc. IPS leve
rages the ubiquitous cell phones equipped
with a number of internal sensors, such as accelerometers, compasses, etc. It attempts to use the
measurements of those sensors to locate the user inside buildings and to create accurate traces of her
motion. We
designed a novel light
weight finite state machine (FSM) to detect the user's steps. We also
employed a hierarchical multi
class support vector machine (SVM) classifier to capture the changing gait
of the user and provide an adaptive stride length. Combi
ning the estimated walking distance with the
direction of motion estimated by the compass and/or the gyroscope, we track the user's motion inside
the building by dead reckoning. In order to restrict the accumulation of error in the user's location, we
roduce a method for error resetting based on ``anchor points" that are typically encountered in
indoor spaces (e.g. stairs, elevators, escalators). Based on the accurate traces constructed, we propose a
method for estimating the building floorplan using

machine learning and computational geometry

We evaluated IPS inside shopping malls and buildings in our campus. Our implementation of IPS using
Android phones provides an accurate step detection with an error of approximately 5.72% regardless

the placement of the phone, 97.74% accuracy of gait type detection and result together in a tracking
error of 6.9%. We also provide a classification tree for detecting the anchor points with 0.2%, 1.3% false
positive and false negative rates respecti
vely. Our proposed error resetting techniques leads to more
than 12 times enhancement in the median distance error.

Conference Paper




UPTIME: Ubiquitous Pedestrian Tracking using Mobile Phones


Moustafa Alzantot, Mo
ustafa Youssef


Inertial navigation; mobile phone tracking; step count estimation; stride length estimation; ubiquitous



The mission of tracking a pedestrian is valuable for many applications including walking distanc
estimation for the purpose of pervasive healthcare, museum and shopping mall guides, and locating
emergency responders. In this paper, we show how accurate and ubiquitous tracking of a pedestrian can
be performed using only the inertial sensors embedded
in his/her mobile phone. Our work depends on
performing deadreckoning to track the user’s movement. The main challenge that needs to be
addressed is handling the noise of the low cost low quality inertial sensors in cell phones. Our proposed
system combine
s two novel contributions: a novel step count estimation technique and a gait
accurate variable step size detection algorithm.

The step count estimation technique is based on a lightweight finite state machine approach that
leverages orientation ind
ependent features. In order to capture the varying stride length of the user,
based on his changing gait, we employ a multi
class hierarchical Support Vector Machine classifier.
Combining the estimated number of steps with the an accurate estimate of the i
ndividual stride length,
we achieve ubiquitous and accurate tracking of a person in indoor environments. We implement our
system on different Android
based phones and compare it to the state
art techniques in indoor
and outdoor testbeds with arbi
trary phone orientation.

Our results in two different testbeds show that we can provide an accurate step count estimation with
an error of 5.72%. In addition, our gait type classifier has an accuracy of 97.74%. This leads to a
combined tracking error of
6.9% while depending only on the inertial sensors and turning off the GPS
sensor completely. This highlights the ability of the system to provide ubiquitous, accurate, and energy
efficient tracking.

Conference Paper


ACM MobiSys 2012


No Need to War
Drive: Unsupervised Indoor Localization


He Wang, Souviek Sen, Ahmed Elgohary, Moustafa Farid, Moustafa Youssef, Romit Roy Choudhury


Location, Mobile phones, Sensing, Landmarks, Recursion


We propose
, an

unsupervised indoor localization scheme

that bypasses the need for war
Our key observation is

that certain locations in an indoor environment present identifiable

signatures on
one or more sensing dimensions. An

elevator, for instance, imposes a
distinct pattern on a smartphone’s

accelerometer; a corridor
corner may overhear a unique

set of WiFi access points; a specific spot may
experience an

unusual magnetic fluctuation. We hypothesize that these kind

of signatures naturally exist
in the environ
ment, and can be

envisioned as internal
of a building. Mobile devices

“sense” these landmarks can recalibrate their locations,

while dead
reckoning schemes can track them

landmarks. Results from 3 different indoor settings, including

a shopping mall, demonstrate
median location errors of

1.69m. War
driving is not necessary, neither are floorplans

the system
imultaneously computes the locations of users

and landmarks, in a manner that they converge

quickly. We believe this
is an unconventional approach to indoor

localization, holding
promise for real
world deployment.

Conference Paper




CrowdInside: Automatic Construction of Indoor Floorplans


Moustafa Alzantot, Moustafa Yousse


shapes, automatic floorplan construction, crowdsourcing


The existence of a worldwide indoor floorplans database can lead to significant growth in location
applications, especially for indoor environments. In this paper, we

present CrowdInside: a
based system for the automatic construction of buildings floorplans. CrowdInside
leverages the smart phones sensors that are ubiquitously available with humans who use a building to
automatically and transparently cons
truct accurate motion traces. These accurate traces are generated
based on a novel technique for reducing the errors in the inertial motion traces by using the points of
interest in the indoor environment, such as elevators and stairs, for error resetting.

The collected traces
are then processed to detect the overall floorplan shape as well as higher level semantics such as
detecting rooms and corridors shapes along with a variety of points of interest in the environment.
Implementation of the system in two

testbeds, using different Android phones, shows that CrowdInside
can detect the points of interest accurately with 0.2% false positive rate and 1.3% false negative rate. In
addition, the proposed error resetting technique leads to more than 12 times enhan
cement in the
median distance error compared to the state
art. Moreover, the detailed floorplan can be
accurately estimated with a relatively small number of traces. This number is amortized over the
number of users of the building. We also discuss
possible extensions to CrowdInside for inferring even
higher level semantics about the discovered floorplans.