Nuzzer algorithm based Human Tracking and Security System for Device-Free Passive System in Smart Phones Environment

Int. J. Advanced Networking and Applications
Volume: 6 Issue: 4 Pages: 2404-2408 (2015) ISSN: 0975-0290
2404
Nuzzer algorithm based Human Tracking
and Security System for Device-Free Passive
System in Smart Phones Environment
R.Ranjani
Assistant Professor, Department of Computer Science and Engineering, IFET College of Engineering, Villupuram.
Email:ranjani2k16@gmail.com
R.Devi
Assistant Professor, Department of Computer Science and Engineering, SKP Institute of Technology, Thiruvannamalai.
Email: r.devi3@gmail.com
M.Kalaimani
Assistant Professor, Department of Computer Science and Engineering, IFET College of Engineering, Villupuram.
Email:mkalaimani@gmail.com
---------------------------------------------------------------------ABSTRACT------------------------------------------------------------
In recent years, majority of researches are focused on localization system for wireless environment. These
researches rely on localization using devices to track the entities. In this paper, we use, a recently proposed
Device-free Passive (DfP) that uses Probabilistic techniques to track locations in large-scale real environment
without the need of carrying devices. The proposed system uses the Access Points (APs) and Monitoring Point
(MPs) that works by monitoring and processing the changes in the received physical signals at one or more
monitoring points to detect changes in the environment. The system uses continuous space estimator to return
multiple location while the mortal is in motion. Our results show that the system can achieve very high probability
of detection and tracking with very few false positives.
Keywords - Device-free Passive (DfP) localization, passive radio map, nuzzer, multiple entity detection.
-------------------------------------------------------------------------------------------------------------------------------------------------
Date of Submission: December 17, 2014 Date of Acceptance: January 24, 2015
-------------------------------------------------------------------------------------------------------------------------------------------------
I. INTRODUCTION
The human tracking system is mainly the localization
process in which the person can send or receive their
location using GPS .The tracking is mainly done by using
the GPS, infrared, ultrasonic, and radio frequency (RF). In
addition, many of these technologies require the device
being tracked to actively participate in the localization
process by running part of the localization algorithm. This
allows the system to provide the user with its location and
other services related to the estimated location. We
introduce the concept in which the localization for
tracking and providing security is mainly done by using
the wireless communication. In the previous technique the
person need to be present within the APs and MPs range
for localization. If there is any disturbance due to the
movement of the person then the tracking cannot take
place and the whole system fails. As the whole the
location was taken as the discrete space estimator so the
accurate result cannot be obtained. The idea is to use
installed wireless data networks to detect changes in the
environment and track the location of entities passively
and without requiring any devices to be attached to these
entities. The DfP concept relies on the fact that RF signals
are affected by changes in the environment, especially for
the frequency ranges of the common wireless data
networks that are currently deployed, such as Wi-Fi, or
envisioned, such as Wi-Max. By placing monitoring
stations that continuously record physical quantities, such
as signal strength or time-of-flight, DfP can analyze these
signals to detect the changes in the environment and
correlate them with entities and their locations. DfP can
be useful in many practical applications including:
intrusion detection and tracking for home and office
applications, which could enhance the safety of law-
enforcement personnel, and low-cost long-range asset
protection, e.g. border protection or protecting railroad
tracks, thereby increasing the value of data networks[1][2].
The DfP system consists of Access point (Aps) and
monitoring point (MPs). The access point is mainly the
global position system (GPS) and the mobile point is the
person themselves. The GPS will get the location of the
person and map it on Google map.
II. RELATED WORK
Many systems over the years have tackled the
localization including the GPS system, infrared, ultrasonic,
and radio frequency (RF).All these systems either require
the tracked entity to be associated with a device that is
being track or requires the device to actively participate in
the localization algorithm. Similar to the DfP concept,
radar-based techniques, computer vision based systems.
However, these systems require specialized hardware and
have a limited range and high installment cost, which
limits their practical aspects. DfP does not require

2405
specialized hardware. In addition it can be considered as
an abstraction for such systems. DfP can be used to
enhance traditional security systems, such as motion
detection and video surveillance by providing non-line-of-
sight detection and lower deployment cost [2]. Although
the concept of DfP looks promising for a number of
applications, there are a number of challenges that need to
be addressed, these challenges include: handling the noisy
physical signal samples to obtain consistent events
recognition and tracking, tracking an entity location, and
identifying an entity. In summary, the DfP localization
concept is unique in detecting, tracking, and identifying
entities without requiring any attached devices or active
participation in the localization algorithms[1]. In addition,
it works with the hardware already installed for the
communication network and thus increases the value of
the network used for localization.
III. TEST BEDS
In this section, we provide actual experimental
results to demonstrate feasibility of the DfP concepts
along with algorithms for solving the detection and
tracking sub functions of the DfP problem. We defer the
discussion of the identification function to the next
section. We start by listing our assumptions and describing
the test bed environment we used to evaluate the proposed
algorithms. To evaluate the proposed techniques, we
conducted two experiments. In general, a zone can be
specified using different techniques. For example, each
room/logical area can be assigned to a zone .Similarly,
zones can be assigned based on the streams passing
through them (Test bed 1).The first experiment was in a
curved corridor with an area of about 24 square meters
.The area is divided into five zones according to the
locations of three APs and four MPs. The second
experiment was performed in a larger area in the second
floor of an office building, where the area (130 square
meters) is divided into six zones. . The RSS for each
stream i is recorded during an offline phase to calculate its
variance, when no person exists in the area. 10
The access points (APs) represent the transmitting units
while the wireless sniffers represent the Monitoring Points
(MPs).
IV. DETECTION
Detection refers to identifying the number of
entities that caused this change. In order to reduce this
ambiguity, and hence enhance the reliability of detection,
we combine different raw data streams to give the overall
system detection. We describe two different statistical
techniques to detect events. The first technique is based on
moving averages and the second is based on the variance.
4.1 MOVING AVERAGE BASED DETECTION
In this technique, events are detected by
comparing two moving averages of the RSSI of a single
stream, with possibly different window sizes. The idea is
to compare the long term signal behavior, which
represents the static environment, to the short term
behavior, which represents the current state, and if there is
a significant change, based on a threshold, an event is
detected. More formally, let qi be a series of raw
measurements over time for one raw data stream3, i.e. a
single MP listening to a single access point. The averages
αl, k and αs,k are defined as follows for time index k:
k+wl-1
a1,k=1/wl ∑qi
i=kb
where wl and ws correspond to the window length for the
two averages αl,kand αs,k respectively. When the relative
difference between the two averages, |αl,k−αs,kαl,k|,
exceeds a threshold (τ ), we declare an event detection for
the time corresponding to t = k + ws. The AS computes
αl,kand αs ,k periodically to check for event detection. As
noted before, the system combines a tunable number of
these alerts (N) from individual streams that give
simultaneous or near simultaneous detections to increase
the overall system performance. The idea is to declare an
event detection if N individual alerts occur within a certain
period. We call this period the alert time buffer (b). Figure
4 shows the effect of the different parameters on
performance. We can see from the figures that, as
expected, increasing either the number of concurrent alerts
used for detection (N) or the detection threshold (τ )
reduces the probability of detection and decreases the
number of false positives. The number of false positives is
insensitive to the time buffer duration (b) and the
probability of detection increases as the time buffer
duration increases. The best values for the parameters are
:( wl= 20, ws= 5,N= 2, τ = 0.03, b = 1). This gives a
1.0probability of detection with zero false positives.
4.2 MOVING VARIANCE BASED DETECTION
The second detection technique is similar to the
first one except that it examines the moving variance of
the raw data and compares it to the variance during the
silence/static period. The detection criterion for any series
qi is based on the variance of the training silence/static
period. For a training period [tstart, tend], we compute the
average of the variances vt and the standard deviation of

2406
the variance σv for the w-sized windows within it. The
moving variance detection technique of the AS detect san
event at time t + w for a single raw stream when vt>vt+
r.σv for an appropriate value of the parameter r. For a
normally distributed variance measurements vt, values
above the threshold will be r standard deviations above the
mean. Again, the system combines a tunable number of
these alerts (N) from individual streams that give
simultaneous or near simultaneous detections, based on a
time buffer (b)to increase the overall system performance.
Figure 5 shows the moving variance applied to the
(MP1,AP2) stream of the raw data in Figure 2. The figure
shows that the system can detect the events, though one
stream may not be adequate as it may lead to false
positives. Figure 6 shows the effect of different parameters
on the performance of the moving variance technique. The
results match the intuition: Increasing the variance
calculation window (w) increases both the probability of
detection and the number of false positives. Increasing
either the number of concurrent alerts used for detection
(N) or the number of standard deviations (r) reduces the
probability of detection and decreases the number of false
positives. The probability of detection is insensitive to the
time buffer duration (b) and the number of false positives
increases as the time buffer duration increases. The best
values for the parameters are :( w = 40,N = 3, r = 4, b = 0).
This gives 1.0 probability of detection with zero false
positives.
V. TRACKING
In order to perform tracking, we need to capture the
relation between signal strength and distance. Since the
relation between signal strength and distance is very
complex in indoor environments, due to the multi-path
effect and other phenomena, we need to capture this
relation at different locations in the area of interest through
a radio map. A radio map is a structure that stores
information about the signal strength at different locations
in the area of interest .This is usually constructed once
during an offline phase. This radio map can be constructed
automatically, based on signal propagation models or can
be constructed manually. We focus in this section on the
manually constructed radio map and defer our discussion
about the automatic construction to the next section.
5.1 PASSIVE RADIO MAP CONSTRUCTION
For the purpose of a DfP system, we call the radio map a
“passive” radio map, in contrast to the “active” radio map
constructed by the current WLAN location determination
systems. Note that the radio map construction process for
a passive radio map is different from the radio
construction process for the traditional active radio map.
For an active radio map. For constructing a passive radio
map, a person stands without carrying any device and the
signal strength characteristics are recorded at the MPs as
received from the APs. The same difference applies to the
automatic construction technique as discussed in the next
section. There are different ways of constructing the radio
map. We use the same method in the Horus active location
determination system where we store at each radio map
point the signal strength histograms from different APs.
For the passive radio map building, we store at each point
of the radio map the signal strength histogram for each
raw data stream, where each data stream corresponds to an
(AP, MP) pair.
5.2 PASSIVE TRACKING
Based on the constructed Radio map, we use a
Bayesian inversion based inference algorithm to compare
the received signal strength vector, which contains an
entry for each raw data stream, to the signal strength at the
different radio map points whose signal strength is stored
in the radio map.
More formally, given a signal strength vector (s) for the
signal strength readings at different MPs from different
APs, with an entry for each pair of (AP, MP), we want to
find the location l in the radio map that maximizes the
probability P (l/s). This can be written as:
arg maxlP(l/s) = argmaxlP(s/l).P(l)P(s)
= argmaxlP(s/l).P (l)
Assuming that all locations are equally probable, the term
P(l) can be factored out from the maximization process.
This leads to argmaxlP(l/s) = argmaxlP(s/l)
Where P(s/l) can be obtained from the constructed radio
map. The two columns represent the case when we used
the first experiment as a training data and the second for
testing and vice versa. The results show that we can obtain
tracking accuracy between 86.3% and 89.7% with simple
algorithms. With an entry for each pair of (AP, MP), we
want to find the location l in the radio map that maximizes
the probability P(l/s). This can be written as:
argmaxlP(l/s) = argmaxlP(s/l).P(l)/P(s)
= argmaxlP(s/l).P (l)
Assuming that all locations are equally probable, the term
P(l) can be factored out from the maximization process.
This leads to argmaxlP(l/s) = argmaxlP(s/l) where P(s/l)
can be obtained from the constructed radio map.
VI. CONTINOUS SPACEESTIMATOR
The continuous space estimator is mainly used to
return the continuous location of the person. In the case of
the discrete space location it returns only the single
location of the person. To get the high accuracy the nuzzer
is used in this. The probability of accurate location is
mainly found by using two techniques. One is spatial
averaging and other is time averaging.
6.1 SPATIAL AVERAGING
This technique is based on treating each location
in the radio map as an object in the physical space whose
weight is equal to the probability assigned by the discrete
space estimator, normalized so that the sum of
probabilities equals one. We then obtain the center of mass
of the k objects with the largest mass, where k is a system

2407
parameter1 <=k <= ||X||. The discrete space estimator 6
will return the location r2, assuming it is the closest in the
signal strength space to the actual person location. Using
the spatial averaging technique, a better location estimate
can be obtained by calculating the center of mass of the
top four locations (k ¼ 4).Note that the estimated location
x need not be one of the
radio map locations.
6.2 TIME AVERAGING
This technique uses a time averaging window to
smooth the resulting location estimates. The technique
obtains the location estimate by averaging the last w
location estimates obtained by either the discrete space
estimator or the spatial averaging estimator. More
formally, given a stream of location estimatesx1; x2; . . . ;
xt, the technique estimates the current location xt at time t
as: tn
Xt= ∑t-min(w,t)+1(xi)
i=t
VII. HANDLING MULTIPLE ENTITIES
The previous section proved the feasibility of tracking a
single entity in the area of interest. In this section, we
address the problem of multiple entities identification and
coarse-grained localization (zone localization).
Identification refers to determining the number of entities
while coarse-grained localization refers to determining the
zones in the area of interest, where the multiple entities are
located. Once this identification and initial localization are
performed, the techniques can be applied to each zone to
determine a more accurate location for the multiple
entities. We describe statistical techniques, based on the
variance of the received signal strength of different
streams, for the identification and coarse-grained
localization of multiple entities. Once the number of
entities, N, is estimated, we need to determine a rough
estimate of where the multiple entities are located. We
recursively apply the relative variance approach to the
streams passing within each zone, where for this purpose a
stream represents the line connecting the AP and MP, to
determine whether an entity is located inside this zone or
not. More formally, given a zone covered by Nz streams,
the average relative variance per zone, RVZoneav , is
calculated as:
!" #$%&'( )
1
+,
- .$% /0
12
234
The N zones with the highest relative variance are
returned as the estimated zones.
7.1 Identifying the Number of Entities
Our approach is based on using the relative
variance, i.e., the ratio between the variance of the human
presence to silence periods as a feature for identifying the
number of entities. The idea is that as the number of
entities in a given area of interest increases, the average
relative variance of all streams covering this area also
increases. More formally, let Ns be the number of streams
covering the entire area. The value of RVav is then
compared against different thresholds to determine the
number of entities. The system declares that there are N
entities in the area of interest if N <RVav< Nþ1, where _N
and _Nþ1 are the Nth and the N þ 1st thresholds.
7.2 Multiple Entities Zone Localization
Once the number of entities, N, is estimated, we
need to determine a rough estimate of where the multiple
entities are located. We recursively apply the relative
variance approach to the streams passing within each zone,
where for this purpose a stream represents the line
connecting the AP and MP, to determine whether an entity
is located inside this zone or not. More formally, given a
zone covered by Nz streams, the average relative variance
per zone, RVZoneav .The N zones with the highest
relative variance are returned as the estimated zones. As a
final note, we point out that localizing entities to within
zones also enhances the scalability of the technique. On
the one side, zones reduce the computational requirements
of the system as it limits the number of radio map CDFs of
zone difference error for localizing multiple entities. A
zone difference of zero means an exact match. Locations
to be processed. On the another side, because each zone is
processed using the streams that mainly have an effect in it
only, this makes the zones virtually independent from each
others. This becomes even more valid as the scale of the
environment increases, as the set of APs and MPs
covering the area would change. Consequently, the
number of sampling points becomes independent from the
number of people as we can calibrate using one person
only. In summary, the scalability is enhanced when the
system uses the multientity detection technique as follows:
1) Identify the number of persons in the environment;
2) Identify the active zones;
3) For each identified active zone person tracking using
the streams that are mainly affected in the identified active
zone.
VIII. DYNAMIC CHANGES IN ENVIRONMENT
Changes in the environment due to the time of day
and day of year, such as temperature and humidity, can
make the operational environment different from the
environment used in constructing the radio map, and hence
may reduce accuracy. Note that for the DfP problem,
changes in people activity patterns should have little effect
on the system, as the system typical environment is at
periods where no one is expected to be in the area of
interest. For environmental changes, based on the
experience from the device-based active localization
systems, their effect is limited. Another possibility is to
use multiple radio maps, corresponding to different

2408
environmental conditions. An automatic passive radio map
construction tool can aid in solving this problem by
capturing the effect of these different changes. Our
experiments, presented, use an independent test set
separate from the one collected during the offline phase.
This quantifies the effect of the dynamic changes in the
environment.
IX. CONCLUSION
We presented the design, implementation, and evaluation
of the Nuzzer DfP localization system. We evaluated the
performance of the Nuzzer algorithm. Our results have
established the proof of feasibility of the DfP concept. In
order to get it to the point that it can be commercially
deployed we have to tackle a number of challenges. We
identified different challenges for realizing the DfP system
and give directions on how to address them. In particular,
we believe that the problem of automatic construction of a
passive radio map is an important challenge that needs to
be addressed for more understanding of the other DfP
challenges and possible solutions. We also believe that
research in the DfP challenges identified in this paper, not
only will have a major impact on the DfP systems, but also
is of value to other localization technologies. Currently,
we are expanding the system in different directions
including: Automatic generation of the passive radio map,
optimizing the APs and MPs positions, and analyzing the
effect of dynamic changes in the environment and
different hardware.
REFERENCES
[1] Moustafa Seifeldin, Ahmed Saeed, Nuzzer: A
Large-Scale Device-Free Passive Localization
System for Wireless Environments, IEEE
Transaction on Mobile Computing, VOL. 12, NO. 7,
JULY 2013.
[2] Wagner.B, Device-free user localization utilizing
artificial neural networks and passive RFID,
Ubiquitous Positioning, Indoor Navigation, and
Location Based Service (UPINLBS), 2012.
[3] Matthew Mah, Challenges: Device-free Passive
Localization for Wireless Environments, Infocom,
Proceedings IEEE, 2012.
[4] M.Ponnavaikko, Location Update Accuracy in
Human Tracking system using Zigbee modules,
(IJCSIS) International Journal of Computer Science
and Information Security, Vol. 6, No. 2, 2009.
[5] M. Youssef and A. Agrawal, The Horus location
determination system, ACM Wireless Networks
(WINET) Journal, 2007.
[6] M. Youssef, M. Mah, and A. Agrawala, Device-free
passive localization, Technical Report CSH860,
UMIACS-TR-2007-13 Department of Computer
Science, University of Maryland, 2007.
[7] M. Youssef, M. Mah, and A. Agrawala, Device-free
passive localization, Technical Report CSH860,
UMIACS-TR, 2007-13
[8] M. Youssef, A. Youssef, C. Rieger, U. Shankar, and
A.Agrawala, Pinpoint: An asynchronous time-based
location determination system, Proceedings IEEE,
June 2006.
[9] Hyuk Lim, Lu-Chuan Kung, Jennifer, Zero-
Configuration, Robust Indoor Localization: Theory
and Experimentation, Proceedings of ACM
Multimedia, 2003.

Nuzzer algorithm based Human Tracking and Security System for Device-Free Passive System in Smart Phones Environment

More Related Content

What's hot (17)

Viewers also liked (11)

Similar to Nuzzer algorithm based Human Tracking and Security System for Device-Free Passive System in Smart Phones Environment (20)

More from Eswar Publications (20)

Recently uploaded (20)

Nuzzer algorithm based Human Tracking and Security System for Device-Free Passive System in Smart Phones Environment