29-09-2016, 10:22 AM
1456728562-IndoorGeolocationScienceandTechnology.pdf (Size: 100.04 KB / Downloads: 8)
ABSTRACT
This article presents an overview of the technical
aspects of the existing technologies for wireless
indoor location systems. The two major
challenges for accurate location finding in indoor
areas are the complexity of radio propagation and
the ad hoc nature of the deployed infrastructure
in these areas. Because of these difficulties a variety
of signaling techniques, overall system architectures,
and location finding algorithms are
emerging for this application. This article provides
a fundamental understanding of the issues
related to indoor geolocation science that are
needed for design and performance evaluation of
emerging indoor geolocation systems.
INTRODUCTION
Recently, there is increasing interest in accurate
location finding techniques and location-based
applications for indoor areas. The Global Positioning
System (GPS) [1] and wireless enhanced 911
(E-911) services [2] also address the issue of location
finding. However, these technologies cannot
provide accurate indoor geolocation, which has its
own independent market and unique technical
challenges. In 1997, while engaged in the Defense
Advanced Research Projects Agency’s (DARPA’s)
Small Unit Operation/Situation Awareness System
(SUO/SAS) program, the lead author of this article
and his research group noticed the need for
fundamental research in accurate indoor geolocation
[3]. The follow-up initiative of the group
attracted the attention of Nokia and other Finnish
organizations to the commercial importance of
indoor geolocation. In recognition of this importance,
an NSF grant was awarded to establish a scientific
foundation in this field.
Accurate indoor geolocation is an important
and novel emerging technology for commercial,
public safety, and military applications [4]. In commercial
applications for residential and nursing
homes there is an increasing need for indoor
geolocation systems to track people with special
needs, the elderly, and children who are away from visual supervision, to navigate the blind, to locate
in-demand portable equipment in hospitals, and to
find specific items in warehouses. In public safety
and military applications, indoor geolocation systems
are needed to track inmates in prisons and
navigating policeman, fire fighters, and soldiers to
complete their missions inside buildings. These
incentives have initiated interest in modeling the
radio channel for indoor geolocation applications
[3, 5], development of new technologies [6], and
emergence of first-generation indoor geolocation
products [7]. To help the growth of this emerging
industry, there is a need to develop a scientific
framework to lay a foundation for design and performance
evaluation of such systems.
Figure 1 illustrates the functional block diagram
of a wireless geolocation system. The main
elements of the system are a number of location
sensing devices that measure metrics related to
the relative position of a mobile terminal (MT)
with respect to a known reference point (RP), a
positioning algorithm that processes metrics
reported by location sensing elements to estimate
the location coordinates of MT, and a display
system that illustrates the location of the
MT to users. The location metrics may indicate
the approximate arrival direction of the signal or
the approximate distance between the MT and
RP. The angle of arrival (AOA) is the common
metric used in direction-based systems. The
received signal strength (RSS), carrier signal
phase of arrival (POA), and time of arrival
(TOA) of the received signal are the metrics
used for estimation of distance. As the measurements
of metrics become less reliable, the complexity
of the position algorithm increases. The
display system can simply show the coordinates
of the MT, or it may identify the relative location
of the MT in the layout of an area. This display
system could be software residing in a
private PC or a mobile locating unit, locally
accessible software in a local area network
(LAN), or a universally accessible service on the
Web. Obviously, as the horizon of accessibility of
the information increases, design of the display
system becomes more complex.
There are two basic approaches to designing a
wireless geolocation system. The first approach is
to develop a signaling system and a network infrastructure
of location sensors focused primarily on
geolocation application. The second approach is to
use an existing wireless network infrastructure to
locate an MT. The advantage of the first approach
is that physical specification, and consequently
quality of the location sensing results, is under control
of the designer. The MT can be designed as a
very small wearable tag or sticker, and the density
of the sensor infrastructure can be adjusted to the
required accuracy of the location finding application.
The advantage of the second approach is that
it avoids expensive and time-consuming deployment
of infrastructure. These systems, however,
need to use more intelligent algorithms to compensate
for the low accuracy of the measured metrics.
Both approaches have their own markets, and
design work on both technologies has been pursued
in the past few years [2, 4, 7, 8].
To develop a scientific foundation, we need to
examine the performance of different signaling
techniques and geolocation approaches. This performance
evaluation needs a suitable model for
radio propagation that reflects the characteristics
of the channel affecting the accuracy of location
sensing and system positioning. In the next three
sections we address technical issues related to
channel modeling, location sensing, and positioning
algorithms for indoor geolocation systems.
CHANNEL CHARACTERISTICS FOR
INDOOR GEOLOCATION
The indoor radio propagation channel is characterized
as site-specific, severe multipath, and low
probability for availability of a line of sight
(LOS) signal propagation path between the
transmitter and receiver [9]. The two major
sources of errors in the measurement of location
metrics in indoor environment are multipath
fading and no LOS (NLOS) conditions due to
shadow fading [3].
Radio propagation channel models are developed
to provide a means to analyze the performance
of a wireless receiver. The performance
criteria for telecommunication and geolocation
systems are quite different [3]. The performance
criterion for telecommunication systems is the bit
error rate (BER) of the received data stream,
while for geolocation systems the performance
measure is the estimated accuracy of location
coordinates. The accuracy of location estimation
is a function of the accuracy of location metrics
and the complexity of positioning algorithms.
Since the metrics for geolocation applications are
AOA, RSS, and TOA, models for geolocation
application must reflect the effects of channel
behavior on the estimated value of these metrics
at the receiver. The existing narrowband indoor
radio channel models designed for telecommunication
applications [9] can be used to analyze the
RSS for geolocation applications. The AOA part
of the emerging 3D channel models developed
for smart antenna applications [10, 11] might be
used for modeling of the AOA for indoor geolocation
applications. However, the existing wideband
indoor multipath channel measurements and models [9] are not suitable for analysis of the
behavior of TOA for geolocation applications.
The existing statistical wideband indoor multipath
models, such as the JTC model [9], represent
multipath characteristics of the channel with
a discrete channel profile similar to the one
shown in Fig. 2. The strength and arrival time of
the paths are so determined that the root mean
square (RMS) delay spread and consequently
BER of a telecommunication receiver obtained
from the simulations using these profile represents
values similar to those obtained from
empirical measurements. If these models are
used for performance evaluation of TOA-based
geolocation systems, the statistics of distance
errors do not reflect the results obtained from
empirical data [3]. Besides, to confirm the modeling
results of a radio channel, empirical measurement
is essential to check the validity of the
model. In the literature there are a number of
measurements of the wideband characteristics of
indoor radio channels for frequencies from 1 to
60 GHz [9]. However, none of these measurements
are useful for geolocation applications
because they do not have a well-calibrated estimate
of the arrival time of the direct LOS
(DLOS) path and a very accurate measurement
of the real physical distance between the transmitter
and receiver [12]. The only available shortrange
measurements calibrated for geolocation
applications are those reported in [12], which are
used in this article to analyze the performance of
super-resolution techniques in the next section.
While we do not have any good models for
the multipath characteristics of indoor radio
channels for geolocation applications, there are
three classes of recent statistical modeling
approaches that can be used to develop reliable
models in the future, which are wideband 2D multipath modeling [3, 5], 3D geometrical statistical
modeling [10], and 3D measurement-based
statistical modeling [11]. In measurement-based
2D statistical modeling, the measurement data
are used to define a multipath profile by
(1)
where Lp is the number of multipath components,
and ak = |ak | e j f k and tk are complex amplitude
and propagation delay of the kth path,
respectively. The strength and statistical characteristics
of the first path and its relative strength
with respect to other paths fit similar results
obtained from empirical data. The measurement
systems for this approach are the same as those
used for telecommunication applications [7, 9].
However, these systems are calibrated for accurate
measurement of the TOA of the DLOS, and
for each measurement the physical distance
between the transmitter and receiver is accurately
recorded. Preliminary measurement and modeling
work in this field is reported in [5, 12]; larger
calibrated measurement databases and more
practical multipath models need further investigation.
In 3D modeling, the mathematical model for
the channels is represented by
(2)
where qk is the AOA of the kth path [11]. While
in 2D modeling each path was associated with a
TOA, in 3D modeling each path is associated with
a TOA and an AOA. The 3D models can be
developed either based on geometric analysis of
the statistics of the paths arriving from different
directions or out of empirical 3D channel measurement
data. The 3D geometrical statistical
models, developed for smart antenna applications,
use an analytical approach to relate propagation
parameters to the structure of scattering in the
environment [10]. In this approach, a mathematical
description of radio propagation based on statistical
building features and a geometric optics
approximation of Maxwell’s equations are
employed to derive relevant radio propagation
models such as distributions of the TOA, AOA,
and RSS. The statistics of the AOA and RSS in
these models can be used directly for indoor
geolocation applications. Further research in this
area is needed to develop statistical models for the
TOA of the DLOS path and its relation to other
paths to make them useful for the analysis of positioning
errors in TOA-based geolocation systems.
In 3D measurement-based statistical modeling,
measured channel characteristics are used to
develop models for AOA, TOA, and RSS. The
major challenge of this approach is the implementation
of a system to measure the 3D characteristics
of the channel. Recently two techniques
have been studied for this purpose. The first
technique mechanically rotates a directional
antenna to measure the strength of the signal
arriving from different directions, and the second
technique measures a set of eight channel impulse responses using an antenna array and
calculates the AOA using signal processing techniques
[11]. Preliminary 3D modeling of an
indoor area using a limited database in a building
is available in [11]. More extensive measurement
and modeling in this field can result in realistic
models for indoor geolocation applications.
LOCATION SENSING TECHNIQUES
As discussed in the introduction, the location sensing
elements measure RSS, AOA, and TOA as
location metrics. The indoor radio channel suffers
from severe multipath propagation and heavy
shadow fading, so the measurements of RSS and
AOA provide less accurate metrics than does TOA
[4]. As a result, similar to GPS systems, independent
systems designed for indoor geolocation normally
employ the more accurate TOA as the
location metric. Systems using existing infrastructures
installed for wireless LANs or the third-generation
(3G) indoor systems may use RSS, AOA,
or less accurate TOA measurements to fully exploit
the existing hardware implementation designed for
traditional telecommunication applications [8]. In
indoor areas, due to obstruction by walls, ceilings,
or other objects, the DLOS propagation path is
not always the strongest; in some cases (e.g.,
NLOS), it may not even be detectable with a specific
receiver implementation [3]. In such cases,
dramatically large errors occur in TOA estimation.
To accurately estimate the TOA in indoor areas,
we need to resort to different and more complex
signaling formats, frequency of operation, and signal
processing techniques that can resolve the
problems. The following subsection is devoted to
accurate TOA estimation techniques.