Content uploaded by Łukasz Szklarski
Author content
All content in this area was uploaded by Łukasz Szklarski on Jan 07, 2024
Content may be subject to copyright.
XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE
Heterogeneous Wireless Sensor Networks Enabled
Situational Awareness Enhancement for Armed
Forces Operating in an Urban Environment
Pawel Kaniewski
Radiocommunications and EW Dept.
Military Communications Institute
Zegrze, Poland
ORCID 0000-0001-8907-8637
Edward Golan
Radiocommunications and EW Dept.
Military Communications Institute
Zegrze, Poland
e.golan@wil.waw.pl
Jan M. Kelner
Institute of Communications Systems
Military University of Technology
Warsaw, Poland
jan.kelner@wat.edu.pl
Emil Guszczyński
ITTI Sp. o.o.
Poznań, Poland
emil.guszczynski@itti.com.pl
Janusz Romanik
Radiocommunications and EW Dept.
Military Communications Institute
Zegrze, Poland
ORCID 0000-0002-6067-423X
Maria D. R-Moreno
TNO
The Hague, The Netherlands
ORCID 0000-0002-7024-0427
Krzysztof Malon
Institute of Communications Systems
Military University of Technology
Warsaw, Poland
krzysztof.malon@wat.edu.pl
Łukasz Szklarski
ITTI Sp. o.o.
Poznań, Poland
lukasz.szklarski@itti.com.pl
Krzysztof Zubel
Radiocommunications and EW Dept.
Military Communications Institute
Zegrze, Poland
ORCID 0000-0002-4207-5786
Paweł Skokowski
Institute of Communications Systems
Military University of Technology
Warsaw, Poland
pawel.skokowski@wat.edu.pl
Krzysztof Maślanka
Institute of Communications Systems
Military University of Technology
Warsaw, Poland
krzysztof.maslanka@wat.edu.pl
RR Venkatesha Prasad
Delft University of Technology
Delft, The Netherlands
r.r.venkateshaprasad@tudelft.nl
Abstract—Situational awareness of armed forces acting in an
urban environment is a key factor determining the success of
military operations. In 2021 the European Defence Agency
established the project on Wireless Sensor Networks for Urban
Local Areas Surveillance. The main goal of the project is to
assess how the situational awareness in an urban environment
can be enhanced with the application of heterogeneous,
autonomous and reconfigurable sensors. The paper presents a
novel comprehensive approach that takes into account
modelling and management of heterogeneous sensors, energy
harvesting techniques, planning and management of the
communication backbone, network security for data transfer
and authorization for secure information exchange. The
architecture of the system and the information flow are
presented. The topology aspects are discussed and the sensing
part is described. The paper finally highlights new essential
enhancements of C2 with particular emphasis on mission
planning, data fusion and threat prediction.
Keywords—situational awareness, wireless sensor networks,
communications management, military operation in urban terrain
I. INTRODUCTION
Armed forces of NATO countries have been engaged in
numerous military operations in various parts of the world.
Therefore combat and peacekeeping operations have been
conducted in different geographic, climatic and cultural
conditions. A significant part of these operations took place
in urban areas often occupied by a combination of non-
combatants and hostile forces [1].
The success of a military operation depends on a number
of factors, including Situational Awareness (SA) that
supports the decision-making process. The SA is often split
in three parts: perception of the elements in the environment,
comprehension of the situation, and projection of future
status. It is obvious that building SA increases the demand for
the efficient delivery of information to the command system
and exchange of information within this system, e.g., data
from sensor networks or reconnaissance systems. Despite
technological developments, i.e., more robust waveforms and
wideband devices offering higher throughput, it is still very
challenging to get reliable and high data rate wireless links in
urban areas.
Besides, limited capacity of batteries powering hand-held
and man-pack radios is another factor influencing
dependability of wireless systems.
One of the main challenges in urban areas is that threats
are more difficult to forecast and can occur almost anywhere,
which means that large areas have to be monitored
persistently. This can be achieved by means of large scale
networks of wireless unattended sensors powered by energy
harvesting sources.
Although many studies on Wireless Sensor Networks
(WSN) have been carried out, they strongly focus on specific
sensor network aspects, e.g. sensor technologies, data fusion
algorithms, or a limited number of nodes in homogenous
terrain. Moreover, these aspects are mostly implemented
during the design phase, which makes the deployed network
very static and thus unable to adapt to changes in the
battlefield. Therefore, what seems to be lacking is a
This work is supported by the European Defence Agency under
CONTRACT NO B 1486 IAP4 GP.
multidisciplinary approach to autonomously adapt a large
network of sensors and energy harvesting resources at
runtime (i.e. after deployment).
To address this problem, in 2021 the European Defence
Agency initiated the project named Wireless sensor Networks
for urban Local Areas Surveillance (WINLAS) [2]. The
objective of the WINLAS project is to demonstrate how
situational awareness of armed forces in an urban
environment can be improved with the application of
heterogeneous, autonomous, reconfigurable sensors.
The rest of the paper is organized as follows: related
works (Section II), urban area scenarios (Section III), system
architecture (Section IV), system components (Section V)
summary and future works (Section VI).
II. RELATED WORKS
WSNs are defined as self-configured and infrastructure-
less wireless networks that enable the observation of physical
or environmental conditions and the direct transfer of
information over the network to the sink where information
is usually visualized and analysed. The application areas of
WSNs are very wide [3].
An overview of defence applications of WSNs was
presented in [4], [5]. The operational context of modern
military engagement has been divided into four scenarios:
battlefield, urban, other-than-war and force protection. These
scenarios were a starting point to define the requirements and
limitations for WSN applications. The types of sensors and
their capabilities determine and limit the use of WSN in
defined scenarios, as presented in Table I.
TABLE I. CLASSES OF MILITARY WSN APPLICATIONS [4]
Sensor
types
Operation scenario
Battle-
field
Urban
OTW
Force
protection
Presence/
Intrusion
SHLM,
AAP
SDT
-
SHLM,
AAP, SDT
CBRNE
RCS
-
VDM
VDM, RCS
Ranging
-
EARS,
INS
BL, INS
EARS, BL,
SDL, PP
Imaging
ASW
SDL,
MCM
-
SDL,
MCM, PP
Noise
-
ATS
ATS
ATS
Abbreviations used in Table I [4]: AAP: Aerostat acoustic payload for transient detection; ASW:
Low-cost acoustic sensors for littoral anti-submarine warf are; ATS: Acoustic threatening sound
recognition system; BL: Time difference of arrival blast localization using a network of disposable
sensors; SDT: Soldier detection and tracking; CBRNE: Chemical, Biological, Radiological, Nuclear
and Explosive; EARS: Early attack reaction sensor; INS: Inertial Navigation System; MCM: Novel
optical sensor system for missile canisters continuous monitoring; OTW: Other Than War; PP:
Perimeter protection; RCS: A low-cost remote chemical sensor for E-UAV platforms; SDL: Sniper
detection and localization; SHLM: Self-healing land mines; VDM: Chemical, biological, and
explosive vapor detection with micro cantilever array sensors.
The performance of the military WSNs application
depends on many factors, including sensor capabilities, type
of sensors, the wireless communication architecture, its range
and appropriate data processing [5], [6].
In [7] authors focused on military requirements for
flexible wireless sensor networks. The following aspects
were taken into account: actual costs per node, the current
mode of deployment (mainly manual network set-up) and
physical size. Their conclusion was, that only limited existing
products meet the current military requirements, as WSNs are
composed of larger sensor devices and consist only of small
numbers of nodes. The authors highlighted that the
dimensions and weight of sensors have to be kept as small as
possible. Moreover, the current trend in the development of
sensors is to get disposable and cheap devices [8] that can be
applied in large quantities. Some types of sensors have to be
able to discover their neighbourhood and to automatically
create the wireless network [14]. Although the sensor
network primarily is considered to be static, it has to be able
to detect and adapt to changes of the topology, e.g., node
disappearance [12]. For most of operations, the WSN is
supposed to cover an area of the size ranging from 5 km2 to
20 km2 [5], [7]. The communication range of a single sensor
should amount to a few hundred meters. Both the number and
the density of sensors will most likely increase significantly
in the future [9]. An important aspect is also the arrangement
of the sensors in a given area. The placement of the sensors
may affect the quality of the results obtained [10], [11].
In literature energy efficiency is noted as one of the key
issues for WSN, because sensor nodes have limited energy
sources (batteries) [15]. Solutions that enable obtaining
energy from the environment to power the sensors are also
proposed [16], [17].
It is assumed that the communication chain will be
composed of sensor nodes, gateway(s) and the sink. Although
bi-directional transfer is supposed, the majority of
information will flow from the sensor node to the gateway.
Data rates depend on the type of sensors and thus may vary
significantly, from low data rates for pressure sensors to high
rates for cameras. Moreover, WSNs are expected to provide
reliable communication resistant to detection and
interception or intentional jamming and interferences from
other WSNs [13].
To achieve the goal of the WINLAS project a novel
comprehensive approach is required, that integrates all of the
following aspects:
• Modelling of various devices for obtaining energy using
e.g. vibration, motion, and piezoelectricity; and
managing network communications with consideration
of different network topologies, energy-dependent
routing strategies, and network traffic load calculations.
• Modelling and management of heterogeneous sensors,
including optical, infrared (IR), radar, life signs, and
chemical, biological, radiological and nuclear (CBRN)
sensors.
• Techniques for combining distributed sensor data for
use in urban environments, including all aspects of
network security for data transfer and authorization for
secure information exchange.
• Networking of a large number of heterogeneous energy
sensing and harvesting devices to provide condensed
information to Command and Control (C2), with
particular emphasis on autonomous adjustment to avoid
human intervention.
III. SCENARIOS
An urban area in the centre of Poland has been chosen to
visualise scenarios and define the topology (Fig. 1). The area
is characterised with the following attributes: a) part of the city
with approx. 12 000 of citizens and an area of approx. 6 km2;
b) the city is surrounded by flat plains to the south and
highlands to the north; c) the river flows through the city from
the north-east to the south-west; d) the city-centre is a dense
urban area with high buildings; e) low and medium-height
buildings prevail in the suburbs; f) an industrial area is located
to the north of the city (power plant, small and medium
factories); g) an airport is located to the south of the city; h)
the nearest village is located by the river at a distance of 3 km
from the city (north-east). The topology has been extended
with further information on weather conditions and available
capabilities for the WINLAS concept. Capabilities include
sensors or sensor nodes, platforms and armed forces. The
capabilities primarily have been based on what is available
within the WINLAS consortium at present. However, the list
could be extended if more would be available.
Fig. 1. WINLAS area for scenarios.
Finally, all scenarios are described as joint missions
involving capabilities from two different countries. The
threats and the available capabilities have been varied between
scenarios. An overview of the scenario visualisations is shown
in Fig. 2.
Four scenarios have been defined:
• a reconnaissance scenario with the aim to confirm
that the urban area is safe to enter (no red forces
present in the area) and in case of a residual threat to
identify, localise and monitor the threat,
• a surveillance scenario with the aim to protect critical
infrastructure and in case of detection of suspicious
activities to track the potential red forces and
intervene, if necessary,
• a patrol scenario with the aim to detect and localise
suspicious activities in an area of interest, to track the
potential red forces and to intervene if necessary,
additionally to extract a person (e.g., evacuate
coalition services' informer),
• a convoy scenario with the aim to escort a
humanitarian aid from the airport to the designated
location in the urban area and in case of suspicious
activities to track the potential red forces and
intervene if necessary.
Fig. 2. Visualisation of the four working scenarios.
IV. SYSTEM ARCHITECTURE
A. Reference Architecture
In literature the typical architecture of the WSN consists
of sensing nodes connected seamlessly through a gateway and
the existing network to the sink. It is assumed that the sink
receives, stores and processes data from sensors. In military
solutions, where the network must be autonomous, the tactical
communication backbone is used as the network infrastructure
and the C2 software plays the role of the sink. Notice that in
WINLAS system we use the name Battlefield Management
System (BMS) that is the extended C2 software.
To get a comprehensive view of the system we considered
its physical and functional architecture. The physical
architecture of the WINLAS system is presented in Fig. 3. The
main components are: BMS, the Communication backbone
and the Sensing component.
The BMS is a software tool composed of the C2 core and
a set of new subcomponents that are application specific.
The Communication backbone is defined as a backbone
network consisting of backbone nodes and gateways with the
tool providing management capabilities.
The Sensing component is created by a set of sensing
devices.
Communication
backbone
BMS
Sensing
component
C2
Automated
Planner & Sensor
Manager
Data store Threat
predictor
Backbone
network Backbone nodes
Gateways
Communication
management
Dynamic Static
Manned system Unmanned system
Managed Unmanaged
WINLAS
system
Fig. 3. System architecture – physical view.
The WINLAS system functional architecture is composed
of the following main components: Information exchange,
Plan and execute and Build situation awareness.
The Information exchange component provides the
following functionalities: a) Authentication; b) Authorization;
c) Data storage; d) Data distribution; e) Communication
management, e.g. setting frequency channels or transmitted
power levels; f) Sensor management, e.g. setting of threshold
values of the sensors; g) Asset control, e.g. autonomous or
remote control of an asset; h) Situation picture distribution.
The Plan and execute component provides the following
functionalities: a) Plan mission, e.g. route/motion/path
planning of platforms and the use of sensors; b) Execute
mission, e.g. sensing and collecting data; c) Monitor mission
plan; d) Re-planning, e.g. new threat identified; e) Monitor
communication backbone; f) Monitor asset/sensor state, e.g.
the level of battery; g) Harvest energy.
The Build situational awareness component provides the
following functionalities: a) Sensing; b) Detection;
c) Classification, e.g. the discrimination between a person, a
car or a truck; d) Identification; e) Localization; f) Tracking;
g) Object fusion; h) Potential threat detection; i) Threat
analysis, e.g. RCIED, saboteurs; j) Threat prediction, e.g.
RCIED explosion, sabotage; k) Situation assessment.
B. Information Flow
Based on the system architecture described above and our
experience in the area of automated command systems, below
we present the presumed information flow for the WINLAS
system. This information flow includes all types of data
exchanged between system elements: a) C2 system data; b)
Sensing data; c) Management data (telemetry data - dynamic
sensors, control and configuration data, status and health
state).
The information flow inside each component of the
WINLAS system and between elements of different
components is shown in Fig. 4.
The C2 component, due to its hierarchical nature,
encompasses the entire system from the high level (HQ) to the
level of the local operator of a single asset or a dismounted
soldier. Depending on the operational scenario lower level C2
instances may have limited functionality, i.e. only C2 Core
function.
Battle
Management System
Communication
backbone
Backbone network
Backbone
nodes GWs
Automated
planner &
sensor manager
Data fusion
& threat
prediction
Communication
management
Sensing
component
Static
Unmanaged
Dynamic
Managed
Manned
ManagedUnmanaged
Unmanned
ManagedUnmanaged
Predicted threats
Updated
situation
awareness
Data from
sensors
Stored data
Current situation
and available
assets
Approved
mission plan/
tasks/requests
Reports,
status of assets
and comm.
backbone
Status Status Tasks/requests
from C2
Requests and config data
Status of
nodes and GWs
Raports and status to C2
Tasks/requests and control/config data from local C2
Status Status Status Status
C2
Data from sensors: detected events, telemetry data
Intel data
Data
storage
C2 Core
Current situation from higher level BMS
C2
Managed
Fig. 4. Information flow.
The information flow between the C2 instances of the
whole WINLAS system is as follows: a) Approved mission
plan (for commanding staff); b) Tasks (for assets); c) Requests
(e.g. to get the asset status or health parameters); d) Situation
awareness; e) Reports; f) Status of assets and communication
backbone.
The information flow within the BMS component includes:
a) Current situation picture - to the Automated Planner &
Sensor Manager; b) Available assets - to the Automated
Planner & Sensor Manager; c) Stored data - from Data storage
to the Data fusion & threat predictor; d) Predicted threat - from
the Data fusion & threat predictor to the Automated Planner
& Sensor Manager; e) Updated situation awareness - from the
Automated Planner & Sensor Manager.
The information flow to the BMS component covers:
a) current situation from higher level; b) C2 data from lower
levels, e.g. reports, status, telemetry data; c) data from sensors
to the Data storage; d) data from INTEL to the Data storage.
The information flow to/from/within the Communication
Backbone component comprises of: a) Tasks / Requests - from
C2; b) Reports / Status - to C2; c) Requests - to Backbone
nodes and GWs; d) Config data - to Backbone nodes and
GWs; e) Status - from Backbone nodes and GWs.
The information flow to/from/within the Sensing
component includes: a) Tasks / Requests - from C2; b) Control
data / Config data - from local C2; c) Status - from assets;
d) Detected events / Telemetry data - to C2/BMS.
V. SYSTEM COMPONENTS
A. Sensing Component
The sensing component consists of different types of
sensors that provide information needed to create and update
the situational picture. A selection of those available within
the WINLAS consortium are described below.
Radiomonitoring sensor will be used for spectrum
situation awareness building [18], [19], [20]. It is a dedicated
device, which means a receiver for spectrum monitoring
placed on a soldier or platform, e.g., an Unmanned Aerial
Vehicle (UAV) or Unmanned Ground Vehicle (UGV).
Software defined radio (SDR) hardware is selected for the
implementation. Energy Detection (Estimated Noise Power –
ED) is proposed as a radio signal activity solution because of
implementation simplicity and low requirements regarding
computational power [21]. The authors suggest to use
emulation (for Data Fusion purposes) of waveform
identification for the Friend, Foe, Neutral (FFN) difference to
support Red Force Tracking (RFT) or Blue Force Tracking
(BFT) functionalities/services. As a hardware platform SDR
USRP B200 mini with frequency range from 70 MHz to 6
GHz and 20 MHz bandwidth was selected. Integration with
UAV platform (DJI Mavic 3 Cine) and microcomputer
(Raspberry Pi4) is proposed for a complete radiomonitoring
sensor (Fig. 5) as a flexible device that can increase mobility
and spectrum monitoring range/coverage area [22].
RADIO
ENVIRONM ENT
PLATFORM (UAV)
COMMUNI CATION
BACKBONE
MICROCOM PUTER
RASPBERRY PI4B
SDR RX
RADIOMONI TORING SENSOR MESSA GES
SPECTRUM S ENSING CONTROL
ENP-ED
DETECTOR
RECEIVED SIGNAL
SPECTRUM MO NITORING
Pd
PfDES
Fig. 5. Radiomonitoring sensor diagram and hardware used for the
development.
From WINLAS architecture point of view the
radiomonitoring sensor as a “Sensing Component” is of the
managed sensor type and can be static or dynamic and manned
or unmanned.
The radiomonitoring sensor will be integrated into WINLAS
with the following parameters: a)frequency to monitor in
[MHz]: single or frequency list, b) bandwidth in [MHz] for
detection (IQ sampling) on the radio frequency: not more than
20 MHz, c) time of measurement [s] for each frequency (might
be the same for the whole list or defined independently), d)
monitoring period [s]: what is the period for the radio
monitoring process, if “0” continuously, e) result type:
hard/soft/IQ samples, desired false alarm selection for hard
and soft detection, f) FFN identification, g) how many last
stored results shall be sent back to the BMS.
The acoustic-seismic sensor combines two detecting
systems. The device is encased in a durable IP55 casing and
has been tested in temperatures ranging from -10 to + 40
Celsius degrees. The sensors' ability to function regardless of
lighting conditions by utilizing sound waves and ground
vibrations is to be of a significant advantage for the system.
The seismic component of the sensor can detect ground
vibrations generated by moving vehicles or people. It is
embedded in the ground to transmit soil vibrations using
acceleration sensors. Its detection algorithms are threshold-
based with noise filtering to reduce external noise, such as rain
droplets.
The detection method of the acoustic component is based
on collecting surrounding noises, filtering them out, then
recognizing and categorizing any unusual sound occurrence.
To make the most of this method, a trained neural network is
constructed to categorize events such as the passing of
wheeled vehicles, general human activity (within a small
radius), and discharged gunshots (larger radius).
Other components inside the sensing subsystem in the
WINLAS project are the CBRN sensors that include:
• Gas Detection Array – Personal (GDA-P), based on Ion
Mobility Spectrometer (IMS) H2O chemistry and
Electrochemical Cell (EC),
• Gas Detection Array – Personal (GDA-P), based on IMS
NH3 chemistry and Photo Ionization Detector (PID),
• AP4C - Flame Photometric Detector (FPD).
The integration of several detection methods enables the
identification of a vast array of chemical compounds. In light
of this, a deployment strategy was implemented that involves
the utilization of two distinct sensor types, namely GDA and
AP4C, as well as two types of GDA devices that employ
different supporting technologies (EC and PID) and are based
on varying chemistry. This approach has been adopted to
augment the system's data filtration capabilities and improve
the precision of identification. The aforementioned sensors are
intended to be integrated into a singular sensor node, which
will collect and refine sensor data prior to transmitting it to
WINLAS' BMS. In order to achieve internal integration a
module called translator was developed. This module is
affixed to individual sensors and facilitates additional wireless
communication with the sensor node. The ESP32-based
module acts as both a gateway between sensor and node
(CBRN sensors) and a primary filtering unit (seismic module).
The fundamental step in the sensor network design is
consideration of the transmission medium. Many of the
Internet of Things (IoT) devices and sensors are majorly
resource-constrained hindering ubiquitous adoption in various
applications, including environmental monitoring, tracking
animals, monitoring some physical parameters of a person,
etc. The current IoT devices are energy constrained since they
use batteries or harvest energy in tiny amounts. They often are
deployed in places where the battery cannot be replaced.
Further, the devices have to be low cost thus, computation
power would be very much limited. Since these devices
cannot have sophisticated protocols running they need to be as
simple as possible. Further, the requirement for the range is
also huge, often multiple kilometres. Any sophisticated MAC
protocol using channel sensing consumes a large amount of
energy and is computationally demanding; this leads to
draining the batteries [27], [28]. Low Power WANs
(LPWANs) can guarantee energy-efficient communication
using hopping techniques. Long-range and efficient
communication requiring minimal amounts of energy for
small payloads is established with a single hop spending not
more than a few hundred micro-watts.
LoRa is a new low-power and long-range communication
protocol. LoRa uses a Chirp Spread Spectrum (CSS)
modulation scheme that is robust towards noise and
interference, see Fig. 6. LoRa offers 50 kbps rate at ranges of
up to 10 km-40 km, depending on the environment
(urban/rural) requiring a maximum of 27 dBm of transmission
power [29]. In contrast, 2G-4G consumes 800x higher power.
LoRa has multiple Spreading Factor (SF) which is the rate at
which the frequency reaches from its lowest value to its
highest. The spreading factor is defined by a value from 7-12.
Fig. 6. CSS Waveform, phase, and instantaneous frequency of LoRa chirps
[30].
A network involving LoRa nodes using a LoRa physical
layer communication is called Long Range Wide Area
Network (LoRaWAN) which offers easy deployment, and
operational longevity to energy-constrained IoT devices. The
IoT devices communicate in a best-effort fashion in extended
ranges. Because of the CSS, the range of the LoRa can be in
terms of kilometers and thus a Gateway can serve a large area.
The deployment, the protocol, and the operation are very
simple since LoRaWAN uses the simple ALOHA-like design
of the MAC layer. The architecture of the LoRaWAN is
shown in Fig. 7. The LoRa-based IoT nodes transmit the LoRa
frames to the Gateways and multiple gateways can receive the
frame. The gateways then decode the frame and send the data
from the frames to the network servers that are accessed by
application servers and then users. In our case, the gateways,
network servers, and applications could be running on the
same system. The single sensors could use LoRaWAN to
directly send information to the command centre.
Fig. 7. LoRa Network Architecture.
End-devices
Gateways
Network server
Application server
real part s(t)
phase φ(t)
inst. freq. f(t)
t (ms)
t (ms)
t (ms)
Phase
Amplitude
Freq.
Robust communication with LoRaWAN is ensured by the
CSS, the MAC being ALOHA-like and multiple gateways
receiving the frames. However, the lack of any central
controller for managing the traffic can results in the packet
collision rate being high in dense LoRaWAN deployments
with high traffic loads.
B. Communication Backbone
Wireless communication in urban areas is always a
challenge for planners of the system and for commanders.
There have been many NATO and EDA projects addressing
this problem [1], [23], [24], [25]. The main conclusion is that
there are no universal solutions fitting all kinds of scenarios.
The organization of the communication backbone always
depends on a given operational goal and the circumstances,
like terrain, opposing forces capabilities or jamming. The
communication system has to be reliable and work efficiently
despite the dynamics and unpredictability of the conducted
operation.
Different types of star-shaped network topologies are
analysed for the WINLAS project, as presented in Fig. 8.
S
Directly
connected
single
sensor
S
Directly
connected
WSN
S
S
S
WSN connected
through dedicated
sensor
S
S
S/G
WSN connected
through gateway
S
S
G
S
S
Many WSNs
connected
through one
gateway
S
G
S
S
S
S
WSN connected
through another WSN
SS
G
S
S
S
S
G
BMS
S
Multihop WSN con nected
through dedicated sensor
S
S
S/G
S
S
G
BMS
Sensor
Gateway
Battle Management System
Fig. 8. Types of the star-shaped topologies.
Directly connected single sensor is the simplest variant of
the WSN network. The main assumption in this variant is that
the sensor uses the same type of radio device for wireless
communication as is used by the receiving centre to
communicate with other elements e.g., armoured vehicles.
Directly connected WSN is a typical variant of the previous
case, where the WSN network is enriched with more sensors.
As a result, a significant increase in the traffic load of a
network can be expected. It can lead to congestion in the
network and problems delivering information to the recipient.
WSN connected through dedicated sensor. When a
dedicated sensor node is in an advantageous location and acts
as an intermediary, it is possible to increase the range of
communication between the WSN network and the receiving
centre. Intermediary nodes must have a memory to store and
forward information. If sensors or a receiving centre move, an
intermediary node may be replaced by another intermediary
node. This can be a part of the planning dynamics dependent
on a change in operational situation.
Multi-hop WSN connected through dedicated sensor
enable a significant increase in the range of wireless
communication, especially in urban operations. The
intermediary nodes in the WSN must have the functionality of
a router and a memory to store and forward information. In
the case of multi-hop topology, a so-called hidden node can
pose problems. It manifests itself as the occurrence of
collisions (interference) in a situation when two sensor nodes
outside the mutual wireless communication range try to
transmit to the same common intermediary node. In the
absence of access control or receipt confirmation mechanisms
the information coming from both sensors will be lost.
WSN connected through gateway is the most popular type
of the communication between WSN and receiving centre
(commander). Most wireless sensors manufacturers provide
their devices with a dedicated hub that communicates with the
sensors via a radio link, while having a wired connection with
the receiving centre.
Many WSNs connected through one gateway. In case of
the same type of radio link in each WSN the gateway node and
sensors should have implemented some collision avoidance
mechanism in the physical layer. In case of different types of
radio links in each WSN the complexity level of the gateway
will grow with the increase in the number of WSN served,
because each WSN will have to have its own transceiver in the
gateway. Additionally, data flows from different WSN will
compete for access to the radio uplink from the gateway to
receiving centre (commander). Some QoS mechanisms should
be implemented in the gateway.
WSN connected through another WSN. This type of
topology allows individual wireless sensor networks to be
connected in a chain, thus significantly increasing the range of
communication with remote sensor networks. However, it
should be remembered that linking networks in a chain leads
to the so-called snowball effect. It manifests itself in an
increase in the amount of information sent on each span of the
chain, leading to overloading of links in the sections closest to
the receiving centre (commander).
In addition to the star-shaped topologies shown above,
mesh topologies were also considered in the project. Mesh
connections between sensors and commander (receiving
centre) enables a significant increase in the area covered by
the many wireless sensor networks. There is also a hidden
node problem as in the multi-hop topology described above.
This type of network requires that individual nodes have
implemented routing mechanisms.
C. Battlefield Management System
Mission Planner
The WINLAS BMS is the system in which information
from the sensor network will be integrated to support the
command and control of the military unit at hand. In the core
of the BMS there is an Automated Planner (AP). Planning is
the area of Artificial Intelligence (AI) that computationally
studies the deliberation process of creating a plan [26]. That
process consists of selecting a sequence of actions that meet
one or more goals and a set of constraints imposed by the
domain. Planning is the reasoning side of acting. When acting
(or executing a process), we need to decide how to perform
the chosen activities while reacting to the environment where
they are taking place. Each action in the plan can be seen as
an abstract task that needs to be refined into sub-actions or
commands that are more concrete.
There are several ways to handle partially observable,
nondeterministic, and unknown environments. We have
sensorless planning (also known as conformant planning) for
environments with no observations; contingency planning for
partially observable and nondeterministic environments; and
online planning and replanning for unknown environments.
This last approach is the one that will be followed by the BMS.
In any of the mentioned cases, it is important to make the
distinction between domain dependent and domain
independent planning. For some type of problems, domain
specific planning methods have been developed that are tailor-
made for that kind of problem. However, for the BMS we have
focused on domain independent planning.
Fig. 9 shows the structure, inputs and outputs of a domain
independent planner. The left part shows the Knowledge
Representation (based on descriptive models) with two inputs:
the domain with the description of the actions based on
preconditions and postconditions, and the problem with the
description of the Initial State (IS) and goals. In the middle
part of the figure is the search process. Before searching for a
solution there is a validation program to syntactically check
that the syntax of the domain and problem files are correct as
well as the logic or procedural inferences. All the algorithms
defined in this part are domain independent to comply with AI
principles. Finally, on the right side, is the solution, that can
be an ordered sequence or parallel actions plan. It may also
not find a solution (failure) based on the domain model
definition and/or the problem provided to the planner and refer
back to the operator.
IS & Goals
(Problem)
Actions
(Domain)
Validator
Sequential Plan
Parallel Plan
Failure
Knowledge
representation
Search domain
independent Result
Fig. 9. Components of an AI automated planner.
Situational Awareness
All the information about the sensor’s status will be
translated into the initial state of the AP as well as a probability
distribution on where for example, the red forces are moving
(this can determine the setting of new goals dynamically),
Fig. 10.
execute mission
monitor resources
build situation
awareness
(re-)plan mission
Mission and sensor management From network
(Prior)
knowledge
Change in asset availability or performance
Change in situation
Fig. 10. Situational awareness function in the WINLAS system.
The state of the preconditions of each action and the
expected effects will be checked to assess whether anything in
the outside world has changed. In case the preconditions or
postconditions do not hold, the AP will replan or explore the
possibility of repairing those parts of the plan that could be
affected by the changes on the preconditions or
postconditions. Then, the user can define the goals or some
goals may be inferred from the information gathered by the
sensors.
VI. SUMMARY AND FUTURE WORKS
The WINLAS project is still in progress. Until now the
following topics were covered: scenarios, requirements for the
system and the reference architecture of the system. The
following tasks are being developed: communications
management, energy harvesting, development and integration
of sensors, development of the mission manager. Below we
shortly refer to each task.
Communications Management
The main goal of communications management is to work
out recommendations for the mission manager. This task
requires extensive simulations and analysis of results. The
following constraints related to the communication backbone
have been identified.
Broadband UHF radios enable high data rate transfer for
the distance up to a few kilometres in LOS conditions. In
urban areas this distance may be significantly reduced, which
may result in unstable or broken communication links
between vehicles. One possible solution to avoid such a
situation is the deployment of a relaying node. Depending on
the scenario, the role of the relaying node can play one of the
recon vehicles that is strategically located or a dedicated
drone. If the drone cannot be launched, e.g., in case of
unfavourable weather conditions, a manpack radio deployed
on a high building may relay the data between vehicles.
Narrowband VHF radios offer voice service or low
throughput data links for the distance up to 25 kilometres in
LOS conditions. Although in urban environment the VHF
radio network may intermittently suffer from short-breaks or
wireless links may be of poor quality, it is considered that
voice services and short messages like BFT will be supported.
Communication with a UAV requires LOS conditions to
get a high throughput data link, e.g. to transfer a videostream
or high quality pictures, and have a reliable control link to
manage the drone, e.g., (re)task, report the battery status or
current flight parameters. In military solutions, a wireless
connection with a tactical drone is provided in one of the
military bands. Currently offered solutions of such wireless
systems cover a wide frequency range from 200 MHz to
6 GHz. Note, that the frequency sub-bands within the band
4.4 GHz – 4.8 GHz are commonly used for this purpose.
In case of wireless sensors, communication links depend
on the distance, local conditions (obstacles) and the type of a
sensor. Videostreams or high quality pictures can be sent from
the camera to the recon vehicle only in LOS conditions. Short
messages such as alerts, alarms or telemetry data can be
exchanged in NLOS conditions, assuming that there is a
dedicated relaying node or that the group of sensors operates
as a WSN with routing mechanisms.
Development of Sensors
During the WINLAS project, a sensor node is developed
whereby individual sensors, connectors, and nodes are
integrated into a unified entity that can function autonomously
and generate data that is not attributable to other system
components. The internal network of the sensor node can be
expanded by incorporating diverse types of sensors. In the
event that an extension of the system is required, the open
design utilizing a robust sensor node and translator modules
as connectors has the capacity to accommodate additional
sensing units.
Development of the Mission Manager
One of the essential preconditions provided to the mission
manager are results of the wireless system simulations which
show constraints in the deployment of sensors and vehicles
from the point of view of the communication manager. The
optimal way to combine this type of data into mission
planning process is being developed.
Other work concerns modelling of the mission with a
description of the initial states and mission goals. In the next
step, a suitable mission plan is created automatically to
achieve the mission goals. The planner selects the appropriate
set of assets, i.e., sensors and platforms for a given plan.
Further works will focus on the development of mechanisms
for automatic assignment and transfer of tasks to individual
sensors and platforms. A UAV equipped with the camera is a
typical example of such a set of platform and sensor that is
foreseen for the first phase of tests. The other types of sensors
will be applied in next phases of the WINLAS project.
ACKNOWLEDGMENT
The authors would like to thank Mrs. Yolanda Rieter-
Barrell for useful advices, suggestions and final proofreading.
REFERENCES
[1] RTO Technical Report TR-IST-067, Technical Communications in
Urban Operations. Report documents the Findings of Task Group IST-
067. Published September 2010.
[2] I. Montiel-Sánchez, “Defence Internet of Things (dIoT). The way
ahead,” European Defence Matters, A magazine of the European
Defence Agency, Issue 14, 2017, p. 15.
[3] L. Gashi, A. Luma, and Y. Januzaj, “The integration of Wireless Sensor
Networks, Mobile Networks and Cloud Engineering for a decision
support system - A Systematic Literature Review,” 2022 International
Congress on Human-Computer Interaction, Optimization and Robotic
Applications (HORA), Ankara, Turkey, 2022, pp. 1-8.
[4] M. P. Durišić, Z. Tafa, G. Dimić, and V.Milutinović “A Survey of
Military Applications of Wireless Sensor Networks”. MECO-2012
Bar, Montenegro.
[5] A. Ali, S.A. Changazi, Y. K. Jadoon, and M. Quasim “Military
Operations: Wireless Sensor Networks based Applications to
Reinforce Future Battlefield Command System”. 2020 IEEE 23rd
International Multitopic Conference (INMIC). Bahawalpur, Pakistan.
[6] S. Pragadeswaran, S. Madhumitha, and S. Gopinath, “Certain
Investigations on Military Applications of Wireless Sensor Networks,”
International Journal of Advanced Research in Science,
Communication and Technology (IJARSCT) Volume 3, Issue 1, March
2021.
[7] M. Winkler, K.-D. Tuchs, K. Hughes, and G. Barclay, “Theoretical and
practical aspects of military wireless sensor networks,” in Journal of
Telecomunications and Information Technology, 2/2008, pp. 37-45.
[8] S. R. Boselin Prabhu, M. Pradeep, and E. Gajendran, “Military
applications of wireless sensor network system,” A Multidisciplinary
Journal of Scientific Research & Education, Volume: 2, Issue: 12,
December 2016.
[9] M. Suchański, P. Kaniewski, J. Romanik, E. Golan, and K. Zubel,
“Radio Environment Maps for Military Cognitive Networks: Density
of Sensor Network vs. Map Quality,” (2019). In book: Cognitive
Radio-Oriented Wireless Networks (pp.195-207).
[10] M. Suchański, P. Kaniewski, J. Romanik, E. Golan, and K Zubel,
“Radio Environment Maps for Military Cognitive Networks:
Deployment of Sensors vs. Map Quality,” In Proceedings of the 2019
International Conference on Military Communications and
Information Systems (ICMCIS), Budva, Montenegro, 14–15 May
2019; pp. 1–6.
[11] K. Junosza-Szaniawski, D. Nogalski, and P. Rzążewski, “Exact and
Approximation Algorithms for Sensor Placement Against DDoS
Attacks,” International Journal of Applied Mathematics and Computer
Science, vol.32, no.1, 2022, pp.35-49.
[12] A. Zhuk, V. Romaniuk, E. Stepanenko, and A. Romaniuk,
“Methodology of the tactical wireless sensor networks control,” 14th
International Conference on Advanced Trends in Radioelectronics,
Telecommunications and Computer Engineering (TCSET), Lviv-
Slavske, Ukraine, 2018, pp. 1047-1050.
[13] S. Laitrakun and E. J. Coyle, “Collision-aware distributed estimation
in WSNs using sensor-censoring random access,” MILCOM 2015
IEEE Military Communications Conference, Tampa, FL, USA, 2015,
pp. 1039-1045.
[14] G. You and Y. Zhu, “Structure and Key Technologies of Wireless
Sensor Network,” Cross Strait Radio Science & Wireless Technology
Conference (CSRSWTC), Fuzhou, China, 2020, pp. 1-2.
[15] N. Mohammad, S. Muhammad, A. Bashar, and M. A. Khan, “Design
and modeling of energy efficient WSN architecture for tactical
applications,” Military Communications and Information Systems
Conference (MilCIS), Canberra, ACT, Australia, 2017, pp. 1-6.
[16] P. Mohan and N. M. R, “Ranking of Sensor Nodes by Optimizing
Sensor Data in Energy Harvesting Wireless Sensor Network,” IEEE
International Conference on Data Science and Information System
(ICDSIS), Hassan, India, 2022, pp. 1-6.
[17] M. T. Le, V. D. Ngo, P. T. Duong, V. T. Tran, and Q. C. Nguyen,
“Dual-Band Funny-Face Wi-Fi Energy Harvester for Indoor Battery-
less Wireless Sensor,” 4th International Conference on Smart Sensors
and Application (ICSSA), Kuala Lumpur, Malaysia, 2022, pp. 1-4.
[18] K. Malon “Dynamic Spectrum Access in Ad-Hoc Radio Networks with
Cognitive Nodes” Military University of Technology: Warsaw,
Poland, 2021; ISBN 978-83-7938-327-6.
[19] P. Skokowski, “Electromagnetic Situation Awareness Building in Ad-
Hoc Networks with Cognitive Nodes”; Military University of
Technology: Warsaw, Poland, 2021; ISBN 978-83-7938-323-8.
[20] P. Skokowski, K. Malon, and J. Łopatka, "Building the electromagnetic
situation awareness in MANET cognitive radio networks for urban
areas", Sensors, vol. 22, no. 3, pp. 716, Jan. 2022
[21] P. Skokowski, K. Malon, and J. Łopatka, "Properties of centralized
cooperative sensing in cognitive radio networks", Proceedings of SPIE
10418 2016 XI Conference on Reconnaissance and Electronic Warfare
Systems (CREWS), vol. 10418, pp. 1041807, Apr. 2017.
[22] K. Malon, P. Skokowski, and J. Lopatka, “Optimization of wireless
sensor network deployment for electromagnetic situation monitoring,”
International Journal of Microwave and Wireless Technologies, 10(7),
pp. 746 – 753, 2018
[23] RTO Technical Report. (February 2009). Command and Control
Challenges in Urban Operations. Final Report of Task Group IST-046
/ RTG-008.
[24] RTO Technical Report 71. (April 2003). Urban operations in the year
2020. Report by the RTO Studies, Analysis and Simulation Panel Study
Group SAS-030.
[25] European Defence Agency. (September 2010). Wireless robust link for
urban force operations. Contract A-0692-RT-GC “WOLF”.
[26] G. Malik, D. Nau, and P. Traverso, (2016). Automated Planning and
Acting. Cambridge University Press.
[27] T. R. Park, T. H. Kim, J. Y. Choi, S. Choi, and W. H. Kwon.
“Throughput and energy consumption analysis of IEEE 802.15. 4
slotted CSMA/CA”. In: Electronics Letters 41.18 (2005), pp. 1017–
1019.
[28] R. Bruno, M. Conti, and E. Gregori. “Optimization of efficiency and
energy consumption in p-persistent CSMA-based wireless LANs”. In:
IEEE Transactions on Mobile Computing 1.1 (2002), pp. 10–31.
[29] Semtech. LoRa® wireless RF products. https://www.semtech.com
/products/wireless-rf/lora-transceivers/sx1261.
[30] Maria Teresa Blanco Abad, Designing and implementing SFMAC: A
MAC protocol for LoRa networks for efficient use of unlicensed bands,
MSc Thesis, Delft University of Technology.
