5G Localization and Context-Awareness

Stefania Bartoletti, Andrea Conti, Davide Dardari, and Andrea Giorgetti

Abstract The fifth generation (5G) wireless ecosystem will be essential for a myriad
of new applications based on accurate location awareness and other contextual
information. Such wireless ecosystem will be enabled by advanced 5G wireless
technologies integrated with existing technologies for the Internet-of-things (IoT)
and the global navigation satellite system (GNSS). First, we will explore the main
5G use cases and key performance indicators (KPIs) presented by the third generation
partnership project (3GPP) where accurate positioning is required. Second, the main
technologies will be described. Then, foundations and signal processing techniques
for accurate localization will be presented. Finally, some context-aware applications
beyond localization are discussed.

1 Localization Use Cases and KPIs

1.1 Use Cases

The 3GPP categorizes the main localization use cases based on verticals, briefly
summarized in the following.

Stefania Bartoletti
CNIT, DE, University of Ferrara, e-mail: stefania.bartoletti@unife.it
Andrea Conti
CNIT, DE, University of Ferrara, e-mail: andrea.conti@unife.it
Davide Dardari
CNIT, DEI, University of Bologna, e-mail: davide.dardari@unibo.it
Andrea Giorgetti
CNIT, DEI, University of Bologna, e-mail: andrea.giorgetti@unibo.it

167
168 Stefania Bartoletti, Andrea Conti, Davide Dardari, and Andrea Giorgetti

1.1.1 Regulatory and Mission Critical

• Accurate localization within an emergency call service refers to the geolocation

of an user when contacting a public safety answering point (PSAP) through a
short dial emergency telephone number (e.g., E-911);
• Accurate localization of a first responder who is injured or incapacitated during
mission critical operations has long been a goal of public safety. A mission critical
service enables a first responder to stay in contact with other first responders as
well as dispatch and command/control during normal and critical operations;
• The alerting service of nearby emergency responders refers to the alert of all
qualified individuals within close vicinity of the victim via their phones with a
request to provide urgent care in case of a medical emergency; The purpose of
this use case is to improve the localization of the emergency responders closest
to the victim, in order to safeguard the quickest availability of care;
• The localization service of emergency equipment outside hospitals refers to the
localization of life-saving medical equipment, such as automated external defib-
rillators (AEDs), deployed throughout public and private spaces in case of need.
This use case is about knowing where the equipment actually is, rather than where
it is supposed to be.

1.1.2 Location-Based Services

• The bike sharing service allows a rider to rent a bike via a mobile app and drop
it off anywhere for the next user. The accurate locations of shared bikes that are
available is required by the riders to find the nearest bike;
• The localization of users is a key service for augmented reality together with the
estimation of motion. Moreover, the access to databases of contextual information
and geo-localized information systems (GIS) need to be provided with low latency;
• The wearable devices such as smart watches can replace mobile terminals to
provide the customer with basic services such as tracking, activity monitoring,
and emergency messages. However, they require higher power durability in order
to replace smart terminals in applications that depend on accurate localization;
• Localization for advertisement push refers to the advertisement the relies on data
analysis of human activity location. For the advertisement to be effective, it needs
to be closely related to the user profile and location in a period of time;
• The location-based flow management refers to the use of location data of people
in public spaces or any transportation hub (airports, metro or rail station, etc.)
facing large passenger flows to elaborate statistics on passengers as well as to
optimise their organisation and signalling to passenger.
5G Localization and Context-Awareness 169

1.1.3 Industry and eHealth

• The localization of persons and medical equipment in hospitals in real time on

the site of the hospitals (indoor and outdoor, even in the presence of large green
areas and several buildings) is important, for example, to notify the medical
staff if patients reach a non-authorised area and to locate caregivers and medical
equipment (e.g. crash cart), especially in emergency situations;
• The localization of patients outside hospitals refers to patients who manage to
leave the hospital without authorisation and ambulatory patients with a potentially
critical condition (e.g. cardiac, diabetes, high-risk pregnancy);
• For a smarter and more efficient waste collection and management, location data
is essential for finding optimal routes for collection vehicles and for locating bins
where sensors detected fires or other anomalies.

1.1.4 Transport (Road, Railway, Maritime, and Aerials)

• Localization for traffic monitoring, management, and control refers to vehicles and
their location in a map of the infrastructure (roads, lanes). The vehicles position
information needs to be managed over multiple road segments and long distance.
This use case addresses a more dynamic implementation, complementing sensor
and videos with position-related data determined using the 5G system;
• Road-user charging (RUC) defines generic services monitoring vehicle positions
(and/or motion) with the aim of levying a charge or a tax based on the way the
road infrastructure is used.
• The tracking of asset and freights has a key role to optimize the overall transporta-
tion efficiency, and to improve end-to-end traceability. Freight tracking enables
more accurate scheduling of all involved operations, while asset tracker should
fulfil very long lifetime (up to 15 years) and position-related data need to be
secured and protected against tampering.
• The accurate positioning of unmanned aerial vehicle (UAV) is important to sup-
port their missions and operations. Each UAV needs to be geo-localized with
high accuracy (absolute position information) to contextualize data collected in
the monitored area (e.g. images of the environment that is flown over).
In specific applications, the network operator can be asked to provide a customized
localization service with different performance for different users. Therefore, the
support of multiple different localization services can be considered as a use case
itself. This can be obtained by relying on multiple technologies for example, 3GPP
technologies and non-3GPP technologies. Different localization methods support
different levels of accuracy capabilities as described in Sec. 2. So, it is suggested
to support negotiation of localization capabilities considering user, application, or
network operator’s demands.
170 Stefania Bartoletti, Andrea Conti, Davide Dardari, and Andrea Giorgetti

1.2 Key Performance Indicators and Key Attributes

The 3GPP introduces several location-based KPIs for 5G applications. The KPIs are
defined based on an absolute or a relative position estimation, which can be further
specialized into an horizontal position (referring to the position in a 2D reference or
horizontal plane) and into a vertical position (referring to the position on the vertical
axis or altitude) [1]. In some applications, the availability of position estimates is
an additional attribute that describes the percentage of time when a positioning
system is able to provide the required position data within the performance targets
or requirements.
In [1], three KPIs are defined for the accuracy of parameter estimation: (1) position
accuracy describes the closeness of the estimated position of the user equipment (UE)
(either of an absolute position or of a relative position) to its true position; (2) speed
accuracy describes the closeness of the estimated magnitude of the UE’s velocity to
the true magnitude; (3) bearing accuracy describes the closeness of the measured
bearing of the UE to its true bearing. For a moving UE, the bearing is a measure of
the velocity’s direction and this KPI can be combined with speed accuracy into the
velocity’s accuracy.
Other three KPI are related to the timing of parameter estimation availability: (1)
latency describes time elapsed between the event that triggers the determination of
the position-related data and their availability at the positioning system interface; (2)
time to first fix (TFF) describes the time elapsed between the event triggering for the
first time the determination of the position-related data and their availability at the
positioning system interface; (3) update rate is the rate at which the position-related
data is generated by the localization system. It is the inverse of the time elapsed
between two successive position-related data.
Moreover, two KPIs are related to the energy consumption for localization: (1)
power consumption indicates the electrical power (usually in mW) used by the
localization system to produce the position-related data.; (2) energy per fix indicates
the electrical energy (usually in mJ per fix) used by the localization system to
produce the position-related data. It represents the integrated power consumption of
the positioning system over the required processing interval, and it considers both the
processing energy and the energy used during the idle state between two successive
productions of position-related data. This KPI can advantageously replace the power
consumption when the positioning system is not active continuously (e.g. device
tracking).
Finally, the system scalability defines the amount of devices for which the posi-
tioning system can determine the position-related data in a given time unit, and/or
for a specific update rate.
Table I summarizes the localization KPIs requirements for positioning use cases
organized per verticals.
5G Localization and Context-Awareness 171

Fig. 1 Potential requirements per use case highlighted by the 3GPP Rel. 16
172 Stefania Bartoletti, Andrea Conti, Davide Dardari, and Andrea Giorgetti

2 Technologies for Cellular Network Localization

In principle, any signal propagating in a wireless environment intrinsically con-

veys position-dependent information that can be exploited for localization. Such a
position-dependent information can be extracted from measurements of signal met-
rics such as received signal strength (RSS), time-of-arrival (TOA), angle-of-arrival
(AOA), phase, or combinations of them, depending on the radio technology. One
or multiple receivers compute signal measurements with respect to one or multi-
ple reference transmitters, and then infer the position by means of a localization
algorithm.
In cellular networks, the reference transmitters can be navigation satellites, cellu-
lar base stations (BSs), and, for cooperative localization, other mobile users. Satellite
navigation has been considered the main technology for localization so far, due to its
global coverage and high accuracy [2]. Cellular-based localization is used as a com-
plementary solution, when there is a lack of satellite visibility due to the blockage
of the satellite signals, which is typically the case in urban and indoor environments.
The positioning methods can be classified into two main categories depending on
the entity that computes the position: (1) mobile-based: the (mobile) device itself
calculates its location by using signal measurements from terrestrial or/and satellite
transmitters. The assistance data from the network can be exploited to perform the
signal measurements and infer the position; (2) network-based: the network location
server infers the position of the mobile device, by means of signal measurements
performed by the network with respect to the mobile device, or signal measurements
performed and sent by the mobile device to the network.
Classical signal processing techniques for cellular networks include
• Trilateration: the position estimate is obtained by intersecting geometric forms,
e.g. circles or hyperbolas, created by distance or angle measurements between
the terminal and the reference transmitters or receivers. Several types of mea-
surements can be used, such as time of arrival (ToA), time difference of arrival
(TDoA), direction or angle of arrival (DoA or AoA), and received signal strength
(RSS).
• Proximity: the known transmitter position is assigned to be the position of the
terminal. An example is the cell-ID method, where the position provided is the
one of the serving base station.
• Fingerprinting: the algorithm is based on finding the best match for a certain
signal measurement, such as RSS, time delay or channel delay spread, from a
database of fingerprints. Each fingerprint is associated with a specific location.
A combination of the previous localization algorithms can be implemented to im-
prove the overall performance, or to support an algorithm that cannot be computed
stand-alone given the lack of signal measurements. More advanced techniques based
on soft information rather than in single value estimates are described in [3]
The choice of the technology and the type of measurements affect the complexity
of the localization process and the KPIs.
5G Localization and Context-Awareness 173

2.1 Localization in 1G - 4.5G Cellular Networks

2.1.1 1G - 3G Cellular Networks

Positioning of UEs exploiting the pervasive diffusion of existing cellular networks,

hence without the necessity to deploy ad hoc and expensive wireless infrastructures,
has been discussed since the introduction of the first generation (1G) for vehicles
location (a comprehnsive survey can be found in [2]). In fact, at that time, UEs
were not equipped with global positioning system (GPS) receivers or, if any, they
were costly. In subsequent years, one of the main driving application was emergency
positioning (e.g., E-911 in U.S.) which became a mandatory requirement of the
U.S. Federal communications commission (FCC). In this context, cellular networks
operators are responsible for positioning UEs with an accuracy of 100 m and 300 m
in 67% and 90% of all positioning attempts, respectively.
Unfortunately, 1G-3G cellular standards were initially designed and optimized
having in mind data and voice communication services but not positioning. Never-
theless, due to the increasingly demand of location-based servicess (LBSs), in 2008
the specifications of positioning methods to be supported in global system for mobile
communications (GSM), also known as 2G, and universal mobile telecommunica-
tions system (UMTS), also known as 3G, were included in the 3GPP standardization
process [4] through the definition of some classes of location services and the in-
troduction of some approaches designed to provide position information using the
available signals structures. They can be classified in UE-based positioning, where
the measurements and position computation are done by the mobile terminal, and
Network-based positioning, where the measurements and position computation are
carried out at network level.
The main methods to estimate the position in cellular networks are:
• Cell-ID: The position of a UE connected to a specific base station (BS), which is
identified by its cell ID, is determined by the location of the BS itself (proximity).
This is a very rough method of positioning which can be improved by considering
the center of gravity of multiple BSs seen by the UE (up to 7 in GSM);
• RSSI: received signal strength indicator (RSSI) measurements can be used to infer
the distance between the UE and the BS. Unfortunately, propagation effects make
the correlation between RSSI and distance weak thus leading to errors in the
order of 150-200 meters. RSSI measurements can be exploited for trilateration or
in fingerprinting techniques, or radiofrequency pattern matching (RFPM), where
RF maps can be created by an advanced radio propagation prediction software,
possibly refined by surveying, and exploited through pattern matching algorithms
to determine the UE position. Fingerprinting can be in principle very accurate
but it requires frequent RF maps updates and it is very sensitive to changes in
propagation conditions;
• Mobile-Assisted time-of-arrival (TOA): In GSM a rough measurement of the sig-
nal round-trip time (timing advance (TA)) is provided in order to synchronize
the UE with the BS timing of slots. When combined with Cell-ID, such mea-
174 Stefania Bartoletti, Andrea Conti, Davide Dardari, and Andrea Giorgetti

Accuracy

1Km
Long-range IoT CID

3G TDOA
100m E-CID
RFPM

10m 4G TDOA
A-GNSS
WLAN/BT

1m RFID Future standards

UWB
10cm
Coverage

indoor outdoor urban outdoor rural

Fig. 2 Qualitative accuracy and coverage of positioning technologies.

surements can slightly increase the positioning performance even though TA is

available only during the call at the UE;
• OTDOA: The observed time difference-of-arrival (OTDOA) is the time-difference
between the system frame numbers (SFNs) generated by two BSs as observed
by the UE. These measurements, together with other information concerning the
position of the involved BSs and the relative time difference (RTD) of the actual
transmission of the downlink signals, is used to estimate the position of the UE.
Since each OTDOA measurement related to a pair of BSs describes a line of
constant TOA difference, yielding a hyperbola in two dimensions, UE position
is determined by the intersection of hyperbolas of at least two pairs of BSs.
Clearly, OTDOA is a UE-based positioning method which requires a specific
implementation at the UE;
• Assisted - GNSS: Cellular network standard protocols have allocated resources
to carry GNSS assistance data to GNSS-enabled mobile devices in both GSM
and UMTS networks. The purpose is to assist the receiver in improving the
performance in terms of startup time, sensitivity and power consumption.
A comparison of different approaches is qualitatively illustrated in Fig. 2.

2.1.2 4G and 4.5G Cellular Networks

The first long-term evolution (LTE) Release 8 did not provide positioning protocols.
3GPP boosted location services in LTE Release 9, delivered in December 2009 [4],
with particular emphasis to emergency calls, as required by FCC E-911. Positioning
methods in LTE networks can be dependent on the radio access technique (RAT),
5G Localization and Context-Awareness 175

Positioning technology
Cell-ID X X

RAT-dependent
E-CID X X X X X
OTDOA X X X X X X
UTDOA X X X X

RFPM X X
A-GNSS X X X X X X X
RAT-independent

TBS X
WLAN X
Bluetooth X
Barometer X

2G 3G 3.9G 4G 4.5G 5G

LTE-A pro (R12)

LTE-A pro (R13)

LTE-A (R11)
LTE-A (R9)
LTE (R8)
GPS

UMTS

Fig. 3 Positioning methods in cellular network standards.

that is making use of LTE signals, or independent of the RAT, that is using other
signals such as GPS.
As can be seen in Fig. 3, most of RAT-dependent positioning methods are similar
to those used in UMTS [5]. E-CID is an improved version of Cell-ID in which cell
ID information is combined with other measurements such as TA, round-trip time,
and angle-of-arrival (AOA). In LTE, OTDOA uses specific downlink signals called
positioning reference signals (PRSs) which are transmitted in certain positioning
subframes of the orthogonal frequency-division multiple access (OFDMA) signal
structure grouped into positioning occasions which occur periodically every 160,
320, 640 or 1280 ms. PRS are received by the UE so that it can perform TOA
measurements (see Fig. 4). The UE measurement is known as the reference signal
time difference measurement (RSTD) which represents the relative time difference
between two BSs. The UE reports its RSTD measurements back to the network,
specifically to the location server, which determines the position of the UE.
In LTE Release 11, the uplink time difference-of-arrival (UTDOA) has been
introduced, thus allowing the network of BSs, also known as eNB in LTE, to collect
time difference-of-arrival (TDOA) measurements of the signal transmitted by the UE
and hence localize it. The UTDOA method is based on network measurements of the
TOA, in at least 3 BSs, of the signal transmitted by the UE. The difference between
two TOAs at two BSs defines a hyperbola and the position of the UE can be calculated
176 Stefania Bartoletti, Andrea Conti, Davide Dardari, and Andrea Giorgetti

BS 1
OTDOA BS1-BS3
hyperbola

PRS
BS 3
PRS

BS 2
PRS

RSTDs

Location RSTDs OTDOA BS2-BS3

Server hyperbola

Core network

Fig. 4 OTDOA-based positioning in LTE.

as in the OTDOA method. The main difference is that now all the processing is done
by the network and no new functionalities need to be implemented in the UE.
FCC recognized that positioning requirements for indoor scenarios cannot be met
by most of operators, thus new requirements were released in 2015. Specifically,
a 50 m horizontal accuracy should be provided for 40, 50, 70, and 80% of emer-
gency calls within 2, 3, 5, and 6 years respectively. For vertical performance, the
operators should propose an accuracy metric within 3 years. In response, most of
RAT-independent positioning methods have been specified in LTE Release 13 (LTE-
Advanced pro, 4.5G) with the purpose to enhance the positioning accuracy, especially
in indoor environments, as required by FCC rules. This was made possible using mul-
tiple different technologies such as wireless local area network (WLAN)/Bluetooth,
barometric pressure sensors (vertical positioning), and terrestrial beacon systems
(PRS beacons and metropolitan beacon systems). Also RAT-dependent methods,
in particular OTDOA, have been enhanced by defining new PRS patterns and PRS
bandwidth extension.

2.2 Localization in 5G Cellular Networks

The main difference between 5G and previous standards is that 5G KPIs requirements
are no longer defined by the regulatory body for emergency calls, but they are driven
5G Localization and Context-Awareness 177

Massive antenna array

Fig. 5 Single-anchor positioning with massive mmWave antenna arrays.

by the 5G use-cases as described in Sec. 1 and are being used in standardization [6].
The KPIs for accuracy, latency, and energy consumption are reported in Sec. 1.2.
The standardization of positioning in 5G is still under discussion within dedicated
task in Release 16 [1]. Localization will be based on the characteristics of the up-
link and down-link signals of new radio (NR) (3GPP-technologies) but also on new
technologies and network configurations, for example, GNSS (e.g. BeiDou, Galileo,
GLONASS, and GPS), Terrestrial Beacon Systems (TBS), Bluetooth, WLAN, RFID,
and sensors [7].
The main breakthrough in 5G is due to the employment of massive multiple-
input–multiple-output (MIMO) beamforming and of millimeter wave (mmWave)
signals. The use of mmWave brings a two-fold advantage: large available bandwidth
and the possibility to pack a large number of antenna elements even in small spaces
(e.g., in a smartphone). Wideband signals offer better time resolution and robustness
to multipath thus improving the performance of OTDOA/UTDOA schemes, as well
as paving the way to new positioning methods such as multipath-assisted localization
exploiting specular multipath components to obtain additional position information
from radio signals [8]. A large number of antenna elements enables massive MIMO
and very accurate beamforming (see Fig. 5). This will make possible the introduction
of single-anchor approaches providing cm-level and degree-level accuracy in 6D
positioning (3D position and 3D orientation) [9], thus overcoming the problem of
deploying a redundant ad-hoc infrastructure which is, nowadays, a major bottleneck
for the widespread adoption of indoor localization systems. In addition, device-to-
device (D2D) are under consideration in Release 16 for ultra-dense networks enabling
cooperative localization, for instance, in vehicle-to-everything (V2X) scenarios [10].
178 Stefania Bartoletti, Andrea Conti, Davide Dardari, and Andrea Giorgetti

2.3 Non-3GPP Technologies for 5G Localization

The lack of service coverage of GNSSs in indoor environments has generated a rich
research activity on the design of indoor localization solutions in the last two decades.
Some solutions exploit acoustic, infrared, laser, inertial, and vision technologies,
whereas others are based on measurements of specific features of radio signals (e.g.,
TOA, RSSI, etc.) [11].
In the context of radio-based positioning technologies, research efforts followed
two main directions: exploitation of existing standards designed only for communi-
cation; and design of ad-hoc standards/solutions for positioning. Recently, particular
emphasis has been given to technologies for IoT applications which typically use
low-cost, low-complexity, and low-energy devices.

2.3.1 Communication-designed Technologies

Several wireless technologies and standards are currently available for WLANs,
wireless sensor networks (WSNs) and IoT applications in general. Examples are Wi-
Fi, radiofrequency identification (RFID), ZigBee and Bluetooth low energy (BLE).
They do not offer specific positioning capabilities, but their transmitted signals can
be exploited to provide different localization performance levels. While RFID and
BLE, due to their limited range, are typically used with proximity methods, Wi-Fi
technology has been successfully adopted in several positioning systems typically
leveraging on fingerprinting methods where meter-level accuracies can be achieved in
many conditions. Wi-Fi ands BLE have already been considered as complementary
technologies in LTE Release 13 to enhance positioning in indoor environments,
especially thanks their wide diffusion.

2.3.2 Ad-hoc Technologies

The most promising ad-hoc technology for high-accuracy positioning in indoor

environment is ultra-wideband (UWB) [12]. This is justified by the fact that the
larger is the signal bandwidth the higher will be the resolution of time measurements
and hence positioning [13, 14].
According to the FCC, an UWB signal is defined as a signal that has fractional
(relative) bandwidth larger than 20% or an absolute bandwidth of at least 500 MHz
[15]. UWB signals are efficiently generated using impulse generators that are simple
and energy efficient. After the FCC allowed the use of UWB signals in 2002 in the
U.S. and the same happened in Europe in 2007 [16] and worldwide, standardization
efforts took place. This effort resulted in the publication of the IEEE 802.15.4a
standard in 2007 [17] as a physical layer (PHY) alternative based on UWB to the
IEEE 802.15.4 standard for WSNs (the PHY of ZigBee). This standard was followed
by the IEEE 802.15.4f standard published in 2012 for applications in the field of port
& marine cargo, automotive, logistics, industrial and manufacturing [18] .
5G Localization and Context-Awareness 179

After an initial slow market penetration, mainly caused by the high-cost of pro-
prietary devices and the fragmented worldwide power emission mask regulations,
since 2014 the market of real time locating systems (RTLS) took off, thanks to the
availability of low-cost chips compliant with the IEEE 802.15.4a standard, and its
growing rate is around 40% yearly, especially in the field of logistic and Industry
4.0. Recently, UWB has been coupled with the RFID technology to detect and track
battery-less tags powered via wireless links [19]. Besides active positioning, thanks
to its peculiarities, the UWB technology enables also other applications like multi-
static radar for non-collaborative localization [20], life signs detection systems, and
through-wall and underground imaging as will be discussed in Sec. 2.3.4.

2.3.3 Long-range IoT Applications

Most of long-range IoT applications (e.g., smart city, asset tracking, smart metering,
smart farming, and smart logistics) are low-rate applications with coverage of tens of
kilometers, and require battery life lasting years (in some cases more than 10 years).
Moreover, nodes have small capabilities in terms of computation and memory, which
makes accurate localization challenging [21]. Currently, two proprietary solutions
are emerging: LoRa and Sigfox. Both have very low throughputs from few tens of
bits per second up to few hundreds of kbps. They are not designed for positioning and
employ narrowband signals that make time measurements very inaccurate because
of the consequent scarce temporal resolution. Despite that, recent studies showed
that rough positioning accuracy in the order of hundred of meters are possible by
properly processing TDOA measurements at BSs [22].
The IoT market is under consideration also by the standardization bodies. The
two main standard technologies for long-range IoT solutions are IEEE 802.11 Long
Range Low Power (LRLP) and the 3GPP narrowband technologies, i.e., LTE-M,
LTE NB-IoT, and EC-GSM-IoT. The positioning capabilities of 3GPP narrowband
technologies were investigated in LTE Release 14. The main positioning algorithms
are enhanced-CID (ECID), OTDOA and UTDOA with a target accuracy of 50 m [23].
The 5G standard is expected to include dedicated protocols for positioning to enable
positioning in IoT applications, even though several technical issues have still to be
studied [24].

2.3.4 Integration with Device-free Localization

An increasing attention is recently being devoted to the capability of detecting and

tracking objects that do not take actions to help the localization infrastructure or that
do not wish to be detected and localized at all. This operation is referred to non-
collaborative localization and has often been undertaken by exploiting a network
of radio sensors able to scan the area of interest through wideband radio signals
to create a radio image of the objects and the environment. These systems are
classified based on whether the network emits a signal designed for target detection
180 Stefania Bartoletti, Andrea Conti, Davide Dardari, and Andrea Giorgetti

In-sensor processing

− Demodulation
− ToA estimation
− Clutter removal
− Ghost mitigation
− CFAR detection
− 1D clustering
− Measurement select.

y Objects trajectory

Network processing

− Detection & localization

− 2D measurement selection
− 2D clustering
− Object tracking
− Target association

1
Fig. 6 A scenario for object/people tracking through a sensor radar network. The processing steps
are performed both locally on sensors and at network level.

and localization (active radar) [25], or the network exploits signals emitted by other
sources of opportunity (passive radar) [26].
Accurate localization via sensor radars becomes particularly challenging in indoor
environments characterized by dense multipath, clutter, signal obstructions (e.g., due
to the presence of walls), and interference. In a real-world scenario measurements
are usually heavily affected by such impairments, severely affecting detection re-
liability and localization accuracy. These operating conditions may be mitigated
by the adoption of waveforms characterized by large bandwidth, e.g., UWB ones
(see Sec. 2.3.2), exploiting prior knowledge of the environment, selecting reliable
measurements, and using various signal processing techniques [27–30]. The UWB
technology, and in particular its impulse radio version characterized by the transmis-
sion of a few nanoseconds duration pulses [31], offers an extraordinary resolution
and localization precision in harsh environments, due to its ability to resolve mul-
tipath and penetrate obstacles. These features, together with the property of being
light-weight, cost-effective, and characterized by low power emissions, have con-
tributed to make UWB an ideal candidate for non-collaborative object detection
in short-range radar sensor networks applications. A sketch of a scenario with the
localization of objects by a radar sensor network is depicted in Figure 6.
Ubiquitous deployment of sensor radar systems integrated with existing commu-
nication infrastructure is expected to open new application scenarios, some of which
have much in common with the use cases of Table 1. For example, through wall
imaging, i.e., the ability to locate indoor moving targets with sensors at a standoff
range outside buildings [32, 33], search and rescue of trapped victims, and people
5G Localization and Context-Awareness 181

counting [34,35] are just a few examples of promising non-collaborative localization

applications.

3 Localization Design in 5G and Beyond

The KPI requirements in 5G scenarios can be fulfilled by exploiting different signal

processing techniques and technologies, which mostly have limited resources for
communication, processing, and memory. The reliability of 5G localization lies in
fusing data and measurements collected from heterogeneous sensors with contextual
information [36] and in designing efficient network operation strategies [37].

3.1 From Foundation to Operation

3.1.1 Fundamental Limits

To provide performance benchmarks and to guide efficient network design and op-
eration, it is important to understand the fundamental limits of localization accuracy
in 5G as well as the corresponding approaches to achieve such accuracy. For this
purpose, the information inequality can be applied to determine a lower bound for the
estimation errors, which is known as the Cramér-Rao lower bound (CRLB), through
the inverse of the Fisher information matrix (FIM) [12].
To evaluate the localization performance in the presence of noise, CRLB-type
performance bounds for the signal metrics under test, e.g., TOA, OTDOA, UTDOA,
RSSI, or AOA are usually considered. Nevertheless, the properties of the signal
metrics depends heavily on the method used to infer user positions, and the use
of certain signal metrics may discard relevant information for localization. Thus, in
deriving the fundamental limits of localization accuracy, it is desirable to fully exploit
the information contained in the received waveforms rather than using specific signal
metrics extracted from the waveforms [12].
Given the complexity of the scenarios considered, the analysis of fundamen-
tal limits for 5G localization should take into account also for multipath and non
line-of-sight (NLOS) propagations which impact localization accuracy especially in
harsh propagation environments (e.g., indoor) [38]. In addition, the case of D2D
cooperation where intranode measurements are available can be analyzed by taking
into account spatial cooperation (together with temporal cooperation in dynamic
scenarios) by characterizing the information evolution in both spatial and temporal
domains [39].
182 Stefania Bartoletti, Andrea Conti, Davide Dardari, and Andrea Giorgetti

3.1.2 Network Operation Strategies

The performance of localization in 5G depends on the transmitting energy, signal

bandwidth, network geometry, and the channel conditions [40, 41]. Such factors
are driven by the network operation strategy, which determines the allocation of
resources, the network nodes from which the measurements are taken, and the
deployment of mobile nodes and base stations. Network operation plays a critical role
in localization since it not only affects the energy consumption, which is a crucial KPI
in 5G applications, but also determines the localization accuracy [37]. For example,
range measurements between two nodes with poor channel conditions consume
significant amounts of energy without improving localization accuracy [28].
Network operation strategies for efficient localization and navigation can be cate-
gorized into several functionalities, including node prioritization (i.e. prioritization
strategies for allocating transmitting resources such as power, bandwidth, and time to
achieve the best trade-off between resource consumption and localization accuracy),
node activation (i.e. activation strategies for determining the nodes that are allowed
to make inter-node measurements so that the localization accuracy of the entire net-
work is maximized), and node deployment (deployment strategies for determining
the positions of new nodes in the network so that the localization accuracy of certain
existing nodes can be maximally improved) [21].

3.2 Advances in Signal Processing

The integration of hybrid technologies fusing measurements from different sensors

such as inertial, GNSS, camera, various RATs, is a challenge for signal processing
in 5G localization. In particular, new localization methods, for instance based on
statistical machine learning, will be needed to achieve accurate, seamless, and robust
localization.

3.2.1 Soft Information

Conventional localization methods rely on single value estimates (SVEs), i.e. each
measurement used for localization corresponds to the estimate of a single-value
metric such as e.g., TOA, OTDOA, UTDOA, RSSI, or AOA, as described in Sec. 2.
Localization accuracy obtained by SVE-based methods depends heavily on the qual-
ity of such SVEs, which degrades in wireless environments, i.e. in the presence of
multipath and NLOS that lead to measurement biases.
To cope with wireless propagation impairments, conventional localization ap-
proaches focus on improving the estimation of single values [42–45]. Techniques to
refine the SVE have been exploited by relying on models for SVEs errors (e.g., the
bias induced by NLOS conditions) [43,46]. Selecting a subset of received waveforms
that contain reliable positional information can also mitigate SVE errors and it can be
5G Localization and Context-Awareness 183

based on features extracted from their samples [28]. In addition, data fusion can be
obtained by considering the SVE of different features as independent or by involving
hybrid models that account for the relationship among different features [47–50]. To
overcome the limitations of SVE, one-stage techniques have been explored that use
measurements to directly obtain positions estimates from the received waveforms
based on a prior model, namely direct positioning (DP) [51–56].
Recently, new localization techniques have been developed that rely on a set of
possible values rather than on single distance estimate (DE), referred to as soft range
information (SRI). To improve the localization performance it is essential to design
localization networks that exploit soft information (SI), such as SRI or soft angle
information (SAI), together with environmental information, such as contextual data
including digital map, dynamic model, and users profiles [3].
The 5G and IoT scenarios offer the possibility to exploit different sensors in the
environments with stringent limitations in terms of energy and power consumption.
In fact, the reliability of multi-sensor IoT lies in fusing data and measurements
collected from heterogeneous sensors with low computation and communication
capabilities [21], and in designing efficient network operation strategies [37]. This
calls for distributed implementation of SI-based localization capable of fusing infor-
mation from multimodal measurements and environmental knowledge. Distributed
localization algorithms require the communication of messages [57–59], which in-
volves high dimensionality depending on the kind of SI. Therefore, it is of utmost
importance to develop SI dimensionality reduction techniques for message passing.

3.2.2 Cooperative localization with D2D communication

D2D communication in 5G are under consideration in Release 16, in particular

for ultra-dense networks enabling cooperative localization, for instance, in V2X
scenarios [10]. Joint spatial and temporal cooperation of devices can yield dramatic
localization performance improvement over conventional approaches since intranode
measurements and mobility (dynamic) models yield new information for localization
and navigation. In particular, joint spatial and temporal cooperation between devices
incurs in associated costs such as additional communication and more complicated
algorithms over the network: 1) the communication among nodes is required for
inter-node measurements and information exchange; and 2) interdependency among
the estimates of the agent positions hinders effective distributed inference algorithms.
In [39], the concept of network localization and navigation (NLN) has been put forth
to exploit spatiotemporal cooperation among nodes. Cooperative algorithms have
been developed based on graphical models, a branch of statistics that makes infer-
ence possible over highly interrelated random variables. Despite the technological
advances in the field, there are still several issues that need to be addressed to realize
accurate, reliable, and efficient NLN.
184 Stefania Bartoletti, Andrea Conti, Davide Dardari, and Andrea Giorgetti

3.3 Beyond Localization

In addition to specific use cases as flow management, traffic monitoring, advertise-

ment push (see Sec. 1), many other applications such as crowd control and detection
of emergency events are important for smart building and public protection. All these
applications rely on ready-to-use analytics that can be defined based on localization
data. The definition of such analytics will primarily leverage basic spatiotemporal
features such as crowd size and people flow. To this aim, both individual and crowd-
centric approaches can be adopted: (i) individual-centric approaches associate the
measured data to single targets/terminals, and run knowledge discovery separately
on each of them; (ii) crowd-centric approaches associate the measured data directly
to a group of users, and run a crowd-level analysis, resulting in lower dimensionality
and complexity [35]. Therefore new methods based on a crowd-centric approach
will be conceived to provide ready-to-use analytics in 5G scenarios and fulfil the
requirements in terms of latency and energy consumption.

References

1. 3GPP, “3GPP TR-2 2.872 : Study on positioning use cases,” in ETSI, Tech. Report 16, 2018.
[Online]. Available: http://www.3gpp.org/
2. J. A. del Peral-Rosado, R. Raulefs, J. A. López-Salcedo, and G. Seco-Granados, “Survey of
cellular mobile radio localization methods: From 1G to 5G,” IEEE Communications Surveys
Tutorials, vol. 20, no. 2, pp. 1124–1148, Secondquarter 2018.
3. S. Mazuelas, A. Conti, J. C. Allen, and M. Z. Win, “Soft range information for network
localization,” IEEE Trans. Signal Process., vol. 66, no. 12, pp. 3155–3168, Jun. 2018.
4. “The 3rd generation partnership project (3gpp).” [Online]. Available: http://www.3gpp.org/
5. S. S. Cherian and A. N. Rudrapatna, “Lte location technologies and delivery solutions,” Bell
Labs Technical Journal, vol. 18, no. 2, pp. 175–194, Sept 2013.
6. 3GPP, “3GPP RP-172795: Handling new SI/WI proposals in RAN,” in ETSI, Tech. Report
12, 2017. [Online]. Available: http://www.3gpp.org/
7. Y. Liu, X. Shi, S. He, and Z. Shi, “Prospective positioning architecture and technologies in
5G networks,” IEEE Network, vol. 31, no. 6, pp. 115–121, November 2017.
8. K. Witrisal, P. Meissner, E. Leitinger, Y. Shen, C. Gustafson, F. Tufvesson, K. Haneda,
D. Dardari, A. F. Molisch, A. Conti, and M. Z. Win, “High-accuracy localization for assisted
living: 5G systems will turn multipath channels from foe to friend,” IEEE Signal Processing
Magazine, vol. 33, no. 2, pp. 59–70, March 2016.
9. A. Guerra, F. Guidi, and D. Dardari, “Single-anchor localization and orientation performance
limits using massive arrays: MIMO vs.beamforming,” IEEE Transactions on Wireless Com-
munications, vol. 17, no. 8, pp. 5241–5255, Aug 2018.
10. H. Wymeersch, G. Seco-Granados, G. Destino, D. Dardari, and F. Tufvesson, “5G mmwave
positioning for vehicular networks,” IEEE Wireless Communications, vol. 24, no. 6, pp. 80–86,
Dec 2017.
11. D. Dardari, E. Falletti, and M. Luise, Satellite and Terrestrial Radio Positioning Techniques -
A signal processing perspective, 1st ed. London, UK: Elsevier Ltd, 2011.
12. Y. Shen and M. Z. Win, “Fundamental limits of wideband localization – Part I: A general
framework,” IEEE Trans. Inf. Theory, vol. 56, no. 10, pp. 4956–4980, Oct. 2010.
13. D. Dardari, A. Conti, U. J. Ferner, A. Giorgetti, and M. Z. Win, “Ranging with ultrawide
bandwidth signals in multipath environments,” Proc. IEEE, vol. 97, no. 2, pp. 404–426, Feb.
2009, special issue on Ultra -Wide Bandwidth (UWB) Technology & Emerging Applications.
5G Localization and Context-Awareness 185

14. S. Bartoletti, W. Dai, A. Conti, and M. Z. Win, “A mathematical model for wideband ranging,”
IEEE J. Sel. Topics Signal Process., vol. 9, no. 2, pp. 216–228, Mar. 2015.
15. First Report and Order 02-48, Federal Communications Commission, Feb. 2002.
16. The Commission of the European Communities, “Commission Decision of 21
February 2007 on allowing the use of the radio spectrum for equipment using
ultra-wideband technology in a harmonised manner in the Community,” Official
Journal of the European Union, 2007/131/EC, Feb. 23, 2007. [Online]. Available:
http://eur-lex.europa.eu/LexUriServ/site/en/oj/2007/l_055/l_05520070223en00330036.pdf
17. “IEEE standard for information technology - telecommunications and information exchange
between systems - local and metropolitan area networks - specific requirement part 15.4:
Wireless medium access control (MAC) and physical layer (PHY) specifications for low-rate
wireless personal area networks (WPANs),” IEEE Std 802.15.4a-2007 (Amendment to IEEE
Std 802.15.4-2006), pp. 1–203, 2007.
18. IEEE Standard for Information Technology - Telecommunications and Information Exchange
Between Systems - Local and Metropolitan Area Networks - Specific Requirements - Part 15.4:
Wireless Medium Access Control (MAC) and Physical Layer (PHY) Specifications for Low
Rate Wireless Personal Area Networks (WPANs) - Amendment: Active RFID System PHY,
IEEE Std 802.15.4f-2012, 2012.
19. A. Costanzo, D. Dardari, J. Aleksandravicius, N. Decarli, M. D. Prete, D. Fabbri, M. Fantuzzi,
A. Guerra, D. Masotti, M. Pizzotti, and A. Romani, “Energy autonomous UWB localization,”
IEEE Journal of Radio Frequency Identification, vol. 1, no. 3, pp. 228–244, Sept 2017.
20. B. Sobhani, E. Paolini, A. Giorgetti, M. Mazzotti, and M. Chiani, “Target tracking for UWB
multistatic radar sensor networks,” IEEE J. Sel. Topics Signal Process., vol. 8, no. 1, pp.
125–136, Feb. 2014.
21. M. Z. Win, F. Meyer, Z. Liu, W. Dai, S. Bartoletti, and A. Conti, “Efficient multi-sensor
localization for the Internet-of-Things,” IEEE Signal Process. Mag., vol. 35, no. 5, pp. 153–
167, Sep. 2018.
22. B. C. Fargas and M. N. Petersen, “GPS-free geolocation using LoRa in low-power WANs,” in
2017 Global Internet of Things Summit (GIoTS), June 2017, pp. 1–6.
23. J. A. del Peral-Rosado, J. A. Lûpez-Salcedo, and G. Seco-Granados, “Impact of frequency-
hopping NB-IoT positioning in 4G and future 5G networks,” in 2017 IEEE International
Conference on Communications Workshops (ICC Workshops), May 2017, pp. 815–820.
24. X. Lin, J. Bergman, F. Gunnarsson, O. Liberg, S. M. Razavi, H. S. Razaghi, H. Rydn, and
Y. Sui, “Positioning for the internet of things: A 3GPP perspective,” IEEE Communications
Magazine, vol. 55, no. 12, pp. 179–185, Dec 2017.
25. M. I. Skolnik, Radar Handbook, 2nd ed. McGraw-Hill Professional, Jan. 1990.
26. F. Colone, D. Pastina, P. Falcone, and P. Lombardo, “WiFi-based passive ISAR for high-
resolution cross-range profiling of moving targets,” IEEE Trans. Geosci. Remote Sens., vol. 52,
no. 6, pp. 3486–3501, Jun. 2014.
27. S. Bartoletti, A. Conti, A. Giorgetti, and M. Z. Win, “Sensor radar networks for indoor
tracking,” IEEE Wireless Commun. Lett., vol. 3, no. 2, pp. 157–160, Apr. 2014.
28. S. Bartoletti, A. Giorgetti, M. Z. Win, and A. Conti, “Blind selection of representative obser-
vations for sensor radar networks,” IEEE Trans. Veh. Technol., vol. 64, no. 4, pp. 1388–1400,
Apr. 2015.
29. S. Savazzi, S. Sigg, M. Nicoli, V. Rampa, S. Kianoush, and U. Spagnolini, “Device-free radio
vision for assisted living: Leveraging wireless channel quality information for human sensing,”
IEEE Signal Process. Mag., vol. 33, no. 2, pp. 45–58, Mar. 2016.
30. M. Chiani, A. Giorgetti, and E. Paolini, “Sensor radar for object tracking,” Proc. IEEE, vol.
106, no. 6, pp. 1022–1041, 2018, (Special Issue on Foundations and Trends in Localization
Technologies).
31. M. Z. Win and R. A. Scholtz, “Impulse radio: How it works,” IEEE Commun. Lett., vol. 2,
no. 2, pp. 36–38, Feb. 1998.
32. A. Yarovoy, L. Ligthart, J. Matuzas, and B. Levitas, “UWB radar for human being detection,”
IEEE Aerosp. Electron. Syst. Mag., vol. 21, no. 3, pp. 10–14, Apr. 2006.
186 Stefania Bartoletti, Andrea Conti, Davide Dardari, and Andrea Giorgetti

33. M. G. Amin, Through-the-wall radar imaging. CRC press, 2016.

34. J. W. Choi, S. H. Cho, Y. S. Kim, N. J. Kim, S. S. Kwon, and J. S. Shim, “A counting sensor for
inbound and outbound people using IR-UWB radar sensors,” in IEEE Sensors Applications
Symp. (SAS), Catania, Italy, Apr. 2016, pp. 1–5.
35. S. Bartoletti, A. Conti, and M. Z. Win, “Device-free counting via wideband signals,” IEEE J.
Sel. Areas Commun., vol. 35, no. 5, pp. 1163–1174, May 2017.
36. A. Conti, S. Mazuelas, S. Bartoletti, M. Z. Win, and W. C. Lindsey, “Soft information for
localization of things,” Proc. IEEE, vol. 107, 2019.
37. M. Z. Win, W. Dai, Y. Shen, G. Chrisikos, and H. V. Poor, “Network operation strategies for
efficient localization and navigation,” Proc. IEEE, vol. 106, no. 7, pp. 1224–1254, Jul. 2018,
special issue on Foundations and Trends in Localization Technologies.
38. K. Witrisal, P. Meissner, E. Leitinger, Y. Shen, C. Gustafson, F. Tufvesson, K. Haneda,
D. Dardari, A. F. Molisch, A. Conti, and M. Z. Win, “High-accuracy localization for assisted
living,” IEEE Signal Process. Mag., vol. 33, no. 2, pp. 59–70, Mar. 2016.
39. M. Z. Win, A. Conti, S. Mazuelas, Y. Shen, W. M. Gifford, D. Dardari, and M. Chiani,
“Network localization and navigation via cooperation,” IEEE Commun. Mag., vol. 49, no. 5,
pp. 56–62, May 2011.
40. A. Conti, M. Guerra, D. Dardari, N. Decarli, and M. Z. Win, “Network experimentation for
cooperative localization,” IEEE J. Sel. Areas Commun., vol. 30, no. 2, pp. 467–475, Feb. 2012.
41. A. Conti, D. Dardari, M. Guerra, L. Mucchi, and M. Z. Win, “Experimental characterization
of diversity navigation,” IEEE Syst. J., vol. 8, no. 1, pp. 115–124, Mar. 2014.
42. I. Güvenç, C.-C. Chong, F. Watanabe, and H. Inamura, “NLOS identification and weighted
least-squares localization for UWB systems using multipath channel statistics,” EURASIP J.
Adv. in Signal Process., vol. 2008, pp. 1–14, Article ID 271 984, 2008.
43. S. Maranò, W. M. Gifford, H. Wymeersch, and M. Z. Win, “NLOS identification and mitigation
for localization based on UWB experimental data,” IEEE J. Sel. Areas Commun., vol. 28, no. 7,
pp. 1026–1035, Sep. 2010.
44. S. Mazuelas, F. A. Lago, J. Blas, A. Bahillo, P. Fernandez, R. Lorenzo, and E. Abril, “Prior
NLOS measurements correction for positioning in cellular wireless networks,” IEEE Trans.
Vehicular Technology, vol. 58, no. 5, pp. 2585–2591, Jun 2009.
45. J. Khodjaev, Y. Park, and A. S. Malik, “Survey on NLOS identification and error mitiga-
tion problems in UWB-based positioning algorithms for dense environments,” in Annals of
Telecommunications, 2009.
46. C. Morelli, M. Nicoli, V. Rampa, and U. Spagnolini, “Hidden Markov models for radio
localization in mixed LOS/NLOS conditions,” IEEE Trans. Signal Process., vol. 55, no. 4, pp.
1525–1542, Apr 2007.
47. J. Prieto, S. Mazuelas, A. Bahillo, P. Fernandez, R. M. Lorenzo, and E. J. Abril, “Adaptive
data fusion for wireless localization in harsh environments,” IEEE Trans. Signal Process.,
vol. 60, no. 4, pp. 1585–1596, Apr. 2012.
48. A. Catovic and Z. Sahinoglu, “The Cramer-Rao bounds of hybrid TOA/RSS and TDOA/RSS
location estimaion schemes,” IEEE Commun. Lett., vol. 8, no. 10, pp. 626–628, Oct. 2004.
49. J.-C. Chen, P. Ting, C.-S. Maa, and J.-T. Chen, “Wireless geolocation with TOA/AOA mea-
surements using factor graph and sum-product algorithm,” in Proc. IEEE Semiannual Veh.
Technol. Conf., vol. 5, Sep. 2004, pp. 3526–3529.
50. P. Deng and P. Fan, “An AOA assisted TOA positioning system,” in Proc. of Int. Conf. on
Commun. Technol., vol. 2, Beijing, China, 2000, pp. 1501–1504.
51. P. Closas, C. Fernández-Prades, and J. A. Fernández-Rubio, “Direct position estimation ap-
proach outperforms conventional two-steps positioning,” in Proc. of European Signal Pro-
cessing Conference (EUSIPCO), Aug. 2009, pp. 1958–1962.
52. P. Closas and A. Gusi-Amigo, “Direct position estimation of GNSS receivers: Analyzing main
results, architectures, enhancements, and challenges,” IEEE Signal Process. Mag., vol. 34,
no. 5, pp. 72–84, Sep. 2017.
53. L. Tzafri and A. J. Weiss, “High-resolution direct position determination using MVDR,” IEEE
Trans. Wireless Commun., vol. 15, no. 9, pp. 6449–6461, Sep. 2016.
5G Localization and Context-Awareness 187

54. N. Vankayalapati, S. Kay, and Q. Ding, “TDOA based direct positioning maximum likelihood
estimator and the Cramér-Rao bound,” IEEE Trans. Aerosp. Electron. Syst., vol. 50, no. 3, pp.
1616–1635, Jul. 2014.
55. J. P. Beaudeau, M. F. Bugallo, and P. M. Djurić, “RSSI-based multi-target tracking by coop-
erative agents using fusion of cross-target information,” IEEE Trans. Signal Process., vol. 63,
no. 19, pp. 5033–5044, Oct. 2015.
56. M. Rydström, L. Reggiani, E. G. Ström, and A. Svensson, “Adapting the ranging algorithm
to the positioning technique in UWB sensor networks,” Wireless Personal Communications,
vol. 47, no. 1, pp. 27–38, Oct. 2008.
57. A. T. Ihler, J. W. Fisher III, R. L. Moses, and A. S. Willsky, “Nonparametric belief propagation
for self-localization of sensor networks,” IEEE J. Sel. Areas Commun., vol. 23, no. 4, pp. 809–
819, Apr. 2005.
58. H. Wymeersch, J. Lien, and M. Z. Win, “Cooperative localization in wireless networks,” Proc.
IEEE, vol. 97, no. 2, pp. 427–450, Feb. 2009, special issue on Ultra -Wide Bandwidth (UWB)
Technology & Emerging Applications.
59. S. Mazuelas, Y. Shen, and M. Z. Win, “Belief condensation filtering,” IEEE Trans. Signal
Process., vol. 61, no. 18, pp. 4403–4415, Sep. 2013.