UAV Aided Disaster Emergency Network

Chapter 1
INTRODUCTION

The emergence of computation offloading rises when the mobile applications are
needed to track on remote servers and to consume energy. The timing necessity of
offloading a task is a challenging one due to their execution time constraints. Mobile
edge server (MEC) should decide when to execute a task and when to offload a task
in order to minimize the energy. UAV is embedded with the on-device camera and
sensors to work for navigation, disaster management and IoT based agricultural
applications. The quality of experience (QoE) has to be ensured between resource
limited devices and MEC server. MEC significantly reduces the latency by avoiding
congestion between transmitted packets and prolonging the UAV battery lifetime for
the QoE. The featuring tasks are computationally offloaded with the aid of deep
reinforcement learning enabled with water strider optimization algorithm. There are
various algorithms related in optimizing the task offloading such as dynamic
partitioning and programming, Lyapunov optimization, Game theoretic approach and
machine learning algorithms. However, the problem is based upon execution time
constraint. Firstly, the locations of UAV are adjusted as per the real-time offloading
methodologies of users. Secondly the trajectory has to be well planned for energy
consumption, maximum throughput. MEC enables UAV to strengthen their coverage
since the channel impairments are the major problem to have line-of-sight links to the
ground users.
* To optimize the UAV edge intelligence based on DRL cooperative methodology
and to allocate minimum power constraint to each UAV.
* To formulate the UAV energy minimization problem as a Markov decision process
to generate the maximum reward and to design edge intelligence algorithm.
* To compute low energy operation with computational resources of UAVs, DRL
enabled MEC framework is proposed in the multi-UAV system for surveillance
report.
* To compare the experimental results with prevailing methodologies refereed in
previous research so as to enable the prominence of the proposed edge
intelligence in UAV.

Dept of ECE, VKIT Page 1

UAV Aided Disaster Emergency Network

Chapter 2

LITERATURE SURVEY
[1] Deep reinforcement learning based computing offloading in unmanned aerial
vehicles for disaster management (2024)

The emergence of Internet of Things enabled with mobile computing has the
applications in the field of unmanned aerial vehicle (UAV) development. The
development of mobile edge computational offloading in UAV is dependent on low
latency applications such as disaster management, Forest fire control and remote
operations. The joint optimization curtails the weighted sum of average energy
consumption and execution delay. This edge intelligence algorithm combined with
DRL network exploits computing operation to increase the probability that at least
one of the tracking and data transmission is usable. The proposed joint optimization
significantly performs well in terms of execution delay, offloading cost and effective
convergence over the prevailing methodologies proposed for UAV development. The
proposed DRL enables the UAV to real-time decisions based on the disaster scenario
and computing resources availability

[2] UAVs for Disaster Management - An Exploratory Review (2023)

The following research article constitutes a comprehensive review of state-of-the-art

methodologies and problems addressed by current academic literature in the field of
deploying autonomous Unmanned Aerial Vehicles (UAVs) for disaster management.
The work compares several papers to identify the current trend of the field. It also
presents an in-depth analysis of possible future research directions in the domain.

[3] Jianwen Xu, Kaoru Ota, and Mianxiong Dong, “Big Data on the Edge Computing
for Disaster Management” Dec. 2020.

After disasters, network communication is highly susceptible to disruption. In this case,

we may need solutions without original architectures to meet the requirements of
connectivity and communication. As a research hotspot, existing studies and practices
in disaster management are often costly and may have to rely on differentiated

Dept of ECE, VKIT Page 2

UAV Aided Disaster Emergency Network

strategies to deal with actual situations. In this paper, we choose UAVs as edge node
carriers and LoRaWAN (Long Range Wide Area Networking) as a communication
method in coping with mobile edge computing (MEC) for disaster management. Here
we propose UAV-mounted MEC task management strategies to achieve emergency
communication enabled by LoRaWAN. The system model includes two parts, air-to-
ground and remote-to-air, in which we choose LoS/NLoS path loss model and
logdistance to describe the connections. The experiment results show that our strategy
can achieve low-cost, long-range MEC service, which can be quickly deployed in the
affected area after disasters. We also choose path loss, SNR (signal-noise ratio), and
channel capacity as performance metrics and prove that our solutions can increase the
channel capacity while maintaining the same level of path loss and SNR .

[4] K. G. Panda, S. Das, D. Sen, and W. Arif, “Design and Deployment of UAV-Aided
Post Disaster Emergency Network” 2019.

Designing a reliable, resilient, and quickly deployable emergency communication

network is a key challenge for post-disaster management. In this paper, a UAV-assisted
emergency WiFi network is proposed to expedite the rescue operations by guiding the
survivors to the nearest rescue camp location. Here, the Raspberry PI (RPI)
development board, mounted on UAV is considered to form a Wi-Fi chain network
over the disaster region. During network set-up, the proposed solutions for the design
challenges like UAV synchronization, avoid communication disruption and
surveillance data management are the key contributions of this paper. The designed
UAV network is capable of doing on-site surveillance and transmitting the data to the
relief center for better rescue planning. One major challenge is to alert a survivor about
the emergency network, which is addressed by designing a captive portal. Furthermore,
to extend the Wi-Fi network, an Android-based application is developed by which each
smartphone acts as a relay for its neighbor. Three types of field experiment are carried
out to evaluate the performance of the designed prototype. It is found from the field
results; the Wi-Fi access point mode and user datagram protocol are more suitable for
network design as compared to AdHoc mode and transmission control protocol,
respectively. It is also observed from the experiment that the maximum hop distance for
the prototype is 280 meters and 290 meters for a Wi-Fi configuration following IEEE
802.11n and IEEE 802.11ac protocol, respectively.

Dept of ECE, VKIT Page 3

UAV Aided Disaster Emergency Network

[5] M. Deryck, J. Wyckmans, W. Joseph, and L. Martens, “Designing UAV aided

emergency networks for large-scale disaster scenarios” 2019.

Today’s wireless communication networks are very reliable. However, in case of a

disaster, these networks can be overwhelmed by a tremendous amount of requests
which they can not cope with. We propose a deployment tool for UAV (unmanned
aerial vehicle)-aided emergency networks for such disaster scenarios. By using UAVs,
femtocell base stations will be brought to and hovered at their assigned location. We
applied this deployment tool on a realistic disaster scenario in the city center of Ghent,
Belgium. The results are very promising although a large amount of drones (> 1000
type 1 or > 370 type 2 drones) is required to provide full coverage for 1 h. Halving the
user coverage results in 1.8 to 2 times less drones. More effectively is to increase the
drone’s fly height. A 10-m higher fly height can result in a reduction up to 13%.
However, above 100 m, the influence is not significant any more. Decreasing the user’s
service level has no significant influence on the number of required drones for the
considered scenario. Furthermore, a prediction model for the number of required drones
based on the intervention duration and the user coverage is proposed.

Dept of ECE, VKIT Page 4

UAV Aided Disaster Emergency Network

Chapter 3

METHODOLOGY

3.1 Problem analysis and discussion

Fig 3.1 UAV edge intelligence system model

MEC-enabled UAVs leverage edge servers to process data in real-time applications.

In this framework, ground mobile Users (GU) receive computing services from many
UAVs with restricted energy B for a predetermined amount of time. Using t = 0, 1,
2, ..., T−1, the operational period is discretized into T times slots, each having a non-
uniform length. Assume that each time slot, or "association between the UAV and
GU," can only have one GU served by the UAV. Only one of M fixed base stations
(BS) may be hovered over by the UAV during each time slot in order to establish a
direct link with the corresponding GU and carry out its offloaded responsibilities.
UAV edge intelligence system model Communication between edge server on people,
vehicles, and embedded sensors in the vicinity is coordinated by catastrophe
communication architecture. The vehicles may have UAVs, even if the employment
of UAVs for ground server deployment is not given explicitly. Distributed and
cooperative sensing enhances the information needed for command and control to
sustain situational awareness.

Dept of ECE, VKIT Page 5

UAV Aided Disaster Emergency Network

Fig 3.2 WiND network model for post-disaster management

Disasters and crises are highly mutable and uncertain situations often resulted in
massive loss of people’s lives and economy. The disaster may be natural or man-
made. The natural disasters Cover earthquakes, cyclones, volcano eruptions, etc.
Designing a UAV based Wi-Fi networks Will be a cost-effective solution due to the
high chance of smartphone availability. Compared to Custom-made relief equipment,
there are no prerequisites for the Wi-Fi network, and it is easy to configure. Thus, Wi-
Fi is the primary technology considered here for designing an emergency Network
known as Wi-Fi Network on Drone (WiND).

3.2 DRL framework for edge intelligence (DRLEI)

The DRL problem is formulated to solve the issues given below. 1. The location
and direction of UAV are difficult to control due to the dynamic environment. The
tasks may arrive and release dynamically so that the task specific requirements
depend on when to execute and when to offload. 2. Even though the conventional
algorithms such as linear and dynamic programming can give the optimal solution
when the number of UAVs are limited, However, the scalability and complexity
raises due to the increase in number of UAVs. 3. The traditional RL optimization
depends on the action specific and reward specific environment. But we proposed the
MDP strategy to learn the new energy efficient task offloading without prior

Dept of ECE, VKIT Page 6

UAV Aided Disaster Emergency Network

knowledge about the dynamic environment. Journal of Electrical Engineering, Vol.

75, No. 2, 2024 97 DRL framework overcomes the limitations and solves scalability
issue and computational complexity issues in order to provide the energy efficient
offloading solution.

Fig 3.3 DRL framework for edge intelligence

A screenshot of the designed captive portal is shown in Through the portal,
necessary information is conveyed to the survivors like the nearest rescue camp
locations, essential first-aid tips, instructions to download Android applications, and
how to use them to convert the smartphone to act as a relay. Through the web page
(captive portal), information such as the number of survivors, their landmarks, and
their health conditions is also collected and saved in a text file. There is a provision
on the web page allowing the survivor to access a specific website to get the
necessary information. The web page is hosted locally from each Raspberry PI (RPI)
rather than in a central location to avoid unnecessary network load. If the commander
of RCC wants to update some information on the web page on-site, then over multi-
hop communication, through a suitable command old web page file will be replaced
by a newly designed web page.

Dept of ECE, VKIT Page 7

UAV Aided Disaster Emergency Network

3.3 MEC using LoRaWAN

Mobile edge computing not only has many technical characteristics inherited from
cloud computing but also puts forward new requirements for equipment requirements
and related technologies. And resource allocation is very critical in affecting
communication and service quality. MEC network system to let mobile users request
radio resources with minimized energy consumption and latency. Focusing on the
non-convex problem of global optimization in computation offloading using a
successive convex approximate method. As a representative of long-range
communication technologies, LoRaWAN can provide low-cost network connections
to geo-distributed IoT devices in urban or rural areas. Researches on LoRa are mainly
about the determination and comparison of the performance in different environments.
Petajajarvi et al. evaluated the signal range and channel attenuation model of
LoWaWAN using 868 MHz (Europe) and 14 dBm as transmission power.

Fig 3.4 UAV-mounted mobile edge computing network model using LoRaWAN.

Here, there is a design of UAV-mounted mobile edge network model using

LoRaWAN In Equation 1, there is a design of a 3-tier network model using MEC
and UAVs. First for User Tier, in an open area R2 there are trapped users who follow
the spatial homogeneous Poisson point process. Suppose the density of users is pu,
which stands for the user number per unit area.

Dept of ECE, VKIT Page 8

UAV Aided Disaster Emergency Network

As a result, the probability of number of users existing in the unit area is then for

……………….. 1

Service Tier, to provide MEC services to User Tier, we need UAVs as carriers to take
edge devices, including Raspberry Pi, and LoRa modules, while patrolling above the
affected area. When any UAV-mounted edge node receives a request from the user, it
will switch to fly mode and draw near the sender by GPS coordinates.

Fig 3.5 (a) Satellite based emergency network, (b)LDRU based emergency network
(c)Ad-Hoc based Emergency network

In a post-disaster scenario, mainly three different strategies are documented for

setting up an emergency communication network. The strategies are based on satellite
communication, Locally Deployed Resource Unit (LDRU), and Ad-Hoc connection.
In Fig.1.1, different types of emergency network based on the above three strategies
are shown. The most common and traditionally used emergency communication
system is satellite-based communication. But, the high cost of satellite phones

Dept of ECE, VKIT Page 9

UAV Aided Disaster Emergency Network

compared to smartphones and the requirement of basic operational Knowledge, limits

its service only to a few survivors and FRs. Use of satellite communication Can be
fully utilized if the smartphone manufacturer incorporates an emergency mode with a
Compatible transmission scheme or by using commercially available products like
bivystick. Both the solutions have an adverse effect on the cost and battery life of the
smartphone.

Emergency communication set-up using LDRU may differ depending upon the
intensity of the disaster. In some cases, only a few cellular Base Stations (BSs) are
damaged Whereas in other cases, almost all the BSs are lost. In Case A, to start
emergency communication, remaining BSs need to change the earlier transmission
scheme to serve a large Coverage area. The remaining area is served by LDRU. For
Case B, specially designed Movable and Deployable Resource Units (MDRUs) are
used. Each MDRU container carries the components such as the wired/wireless
transceiver, switches/routers, servers, power source unit, and air conditioner for
providing a communication service. It is transported to the disaster zones Through
helicopters, trucks, airships, etc. Although MDRUs have quite advanced features such
as remote operation and maintenance, modularized and virtualized MDRU functions,
but it requires a few days for restoration. The golden time for rescue may be missed if
we solely rely

On MDRU based communication. With the above considerations, wireless Ad-

Hoc network has been considered as an appealing Technology for emergency
communication. Due to no pre-existing infrastructure requirement For Ad-Hoc
network set-up, this technology has garnered major attention as a solution for
Designing a post-disaster communication system. The basic idea behind the Ad-Hoc
network is to set-up a temporary multihop communication link between two or
several nodes where each Node acts a router and host at the same time. With
evolution, new Ad-Hoc paradigms are Developed for different disaster scenarios such
as a Mobile Ad-Hoc NETwork (MANET), Vehicular Ad-Hoc NETwork
(VANET),Delay Tolerant Network (DTN), Wireless Sensor Network (WSN),
Professional/Private Mobile Radio (PMR) and Wireless Mesh Network (WMN).Over
a disaster region, MANETs can be formed by using two main wireless Technologies,
namely, Wi-Fi and Bluetooth. Every smartphone and laptop acts as a node for the
MANET and the node mobility is considered to be low. To implement MANET,
Dept of ECE, VKIT Page 10
UAV Aided Disaster Emergency Network

Network Autoconfiguration Software (NAS) and all the proactive and reactive
routing protocols must be Installed in each smartphone and laptop in advance.
VANET and DTN are two special cases of MANET. In VANET, the node mobility is
considered to be high (vehicular speed). VANET is Useful in a disaster scenario such
as massive traffic accidents, terrorism manifest, and Land slides,etc. In Rescue ME
system is developed over VANET to store user location securely and routinely using
existing infrastructure and use it during disaster rescue. DTN is useful when the
density of nodes over the region is low. In that case, MANET will be malfunctioning
in route Discovery and route establishment. DTN can be applied for low and high
mobility case. MANETs have many challenges such as prediction of accurate
mobility model , design of Multihop energy efficient routing protocol fit for different
node topology, network security issue Such as nasty neighbor relaying packets.

3.4 Results and discussion

 Simulation environment used
The list of simulation parameters. The fundamental parameters for UAV are
frequency, CPU cycles, UAV transmitted power, flying power and hovering power
etc. The evaluated parameters are simulated through MATLAB software using
Laptop core i3 with 16 GB RAM and 1 TB ROM.

 Discussion
The evaluation parameters of the proposed algorithm are velocity, Average
battery Energy of the UAV, computation system delay and average throughput. These
parameters are compared with the existing algorithms such as Q-Learning which is a
popular method in this MEC enabled UAV, The UAVs features, parameters, and
implementation method (simulation) were all completed under identical
circumstances and using the simulation parameters. The simulation environment
consists of K UAVs, where K varies from 2 to 12 for better computational complexity.

Dept of ECE, VKIT Page 11

UAV Aided Disaster Emergency Network

Fig 3.6. Average reward of UAV

Fig 3.7 UAV battery energy level

Dept of ECE, VKIT Page 12

UAV Aided Disaster Emergency Network

Battery energy level is maintained as a number of rounds. Figure 8 displays the

simulation and comparison outcomes of the aforementioned techniques in terms of
the battery lifespan of UAV networks. The proposed algorithm convergence analysis
is shown with the increasing sample duration. The DRL algorithm tends to avoid
dropping tasks by prolonging the average completion time in order to achieve a
minimum execution cost. indicates that the average energy consumption of UAV
while using DRLEI algorithm. The comparative parameters are average execution
cost and with computational capacity and varying with offloading task with
prevailing algorithms proposed in related works. It was discovered that the average
total cost in terms of computing power and energy consumption are sufficient to
demonstrate the effectiveness of the proposed RL-based edge computing algorithm.
The proposed DRLEI reduced the offloading cost as well as average execution cost
by 52.13%, 43.5% and 28.7% in terms of computational cost, task size and the drop.
clearly examined the effect of the proposed DRLEI in comparison with DQN, local
and edge computing.

Dept of ECE, VKIT Page 13

 UAV Aided Disaster Emergency Network

Chapter 4

Applications
1. Search and Rescue: UAVs can be equipped with cameras, thermal sensors, and other
imaging technologies to search for and locate survivors in disaster-hit areas. UAVs
can cover large areas quickly and provide rescuers with real-time information,
making the search and rescue process faster and more efficient.

Fig 3.8 Dammaged area assessment using UAV

2. Damage Assessment: UAVs can be used to assess the extent of damage caused by a
natural disaster. By capturing aerial images and videos of affected areas, UAVs can
help identify areas that need immediate attention, prioritize relief efforts, and estimate
the cost of damage.
3. Delivery of Essential Supplies: UAVs can be used to deliver essential supplies such
as food, water, and medical aid to people in disaster-hit areas. UAVs can reach areas
that are inaccessible to ground vehicles, making it possible to provide aid to people
who need it urgently.

Dept of ECE, VKIT Page 14

 UAV Aided Disaster Emergency Network

Fig 3.8 Dammaged area assessment using UAV

4. Communication and Connectivity: UAVs can be used to establish temporary
communication networks in disaster-hit areas, providing rescuers with a means of
communication even when traditional communication networks are down. UAVs can
also be used to provide Internet access and other communication services to people in
disasterhit areas.

Dept of ECE, VKIT Page 15

 UAV Aided Disaster Emergency Network

5. Mapping and Planning: UAVs can be used to create 3D maps of disaster-hit areas,
providing rescuers with a comprehensive view of the affected area. These maps can
be used to plan relief efforts, allocate resources, and make informed decisions about
the best way to provide aid to people in need.
6. Overall, UAVs have numerous applications in post-disaster emergency networks, and
their use can greatly improve the speed and efficiency of disaster response and relief
efforts, potentially saving lives and reducing the impact of natural disasters.

Dept of ECE, VKIT Page 16

 UAV Aided Disaster Emergency Network

4.1 Advantages

1. Quick response time: UAVs (unmanned aerial vehicles) can be deployed quickly and
reach the disaster zone much faster than human responders, enabling faster response
times.
2. Wide coverage area: UAVs can cover a large area in a short amount of time,
providing an overview of the disaster zone and helping responders to identify areas
that need immediate attention.
3. Remote sensing and monitoring: UAVs equipped with cameras and sensors can
provide real-time information about the disaster area, including the extent of damage,
the location of survivors, and the presence of hazards such as fires or floods.
4. Communication support: UAVs can be used to establish communication networks in
disaster areas, providing critical connectivity to emergency responders and survivors.
5. Reduced risk to responders: The use of UAVs in disaster response can reduce the risk
to human responders by providing an aerial view of the disaster area and enabling
responders to assess the situation before entering.

4.2 Disadvantages

1. Limited battery life: UAVs have a limited battery life and require frequent recharging
or replacement of batteries, which can be a challenge in remote disaster zones.
2. Limited payload capacity: UAVs have limited payload capacity, which can limit the
amount of equipment or supplies that can be transported to the disaster zone.
3. Weather conditions: Weather conditions such as strong winds, rain, or fog can limit
the ability of UAVs to operate effectively, which can delay emergency response
efforts.
4. Data management: The large amount of data generated by UAVs can be difficult to
manage, process, and analyze, requiring specialized software and expertise.
5. Privacy concerns: The use of UAVs in disaster response raises privacy concerns,
particularly if they are equipped with cameras or sensors that capture images or data
of individuals or private property without consent.

Dept of ECE, VKIT Page 17

UAV Aided Disaster Emergency Network

Chapter 5

CONCLUSION
The proposed DRLEI strategy is implemented with the proven results when
offloading computation. The facilitation of offloading is done through the ground
edge server which helps the edge users for computationintensive activities, The
successful completion of all execution and offloading tasks based on energy
consumption and task execution latency. The DRLEI framework helps in cost
optimization and computational power optimization as weighted sum average. An
agent performs best optimization through rigorous training phase and deciding the
best offloading strategy and taking actions regarded with the novel reward functions
of the proposed DRLEI scheme. Lastly, the convergence of the DRLEI is tested
through simulation. In comparison with DQN, edge, and local execution strategies.
The comparative results are outperformed and remarkable. Reduced operational
ranges, lower payloads, and shorter flight periods are outcomes of singleuse UAV
restrictions that could be solved by the proposed work combining networked and
collaborative UAVs. System setup took less than five minutes in a large-scale fire
practice, and it only took a few more minutes to provide aerial surveillance of the
whole region.

Dept of ECE, VKIT Page 18

 UAV Aided Disaster Emergency Network

Chapter 6

FUTURE SCOPE
1. The use of Unmanned Aerial Vehicles (UAVs) in post-disaster emergency networks
has the potential to greatly improve response times and efficiency in disaster relief
efforts. There are several areas where UAVs can be employed, including search and
rescue, damage assessment, and delivery of essential supplies.
2. In the future, we can expect to see further advancements in UAV technology that will
enable them to perform more complex tasks and operate for longer periods of time.
For example, improvements in battery technology may allow UAVs to stay in the air
for extended periods, while advances in sensors and imaging technology could enable
more precise damage assessment and identification of survivors.
3. Furthermore, developments in AI and machine learning could help UAVs
autonomously identify and locate survivors and deliver aid packages to those in need.
This would reduce the workload on human operators and allow for faster response
times.
4. In addition, the use of UAVs in disaster response could also lead to the development
of more resilient communication networks. UAVs could be equipped with
communication relays that can provide connectivity to remote or inaccessible areas,
allowing rescuers to coordinate their efforts more effectively.
5. Overall, the future scope for UAVs in post-disaster emergency networks is very
promising. As technology continues to improve, we can expect to see UAVs playing
an increasingly important role in disaster response efforts, helping to save lives and
reduce the impact of natural disasters.

Dept of ECE, VKIT Page 19

UAV Aided Disaster Emergency Network

REFERENCE

[1] Deep reinforcement learning based computing offloading in unmanned aerial vehicles for disaster
management - Anuratha Kesavan1, Nandhini Jembu Mohanram Soshya Joshi ,Uma Sankar (2024)
https://doi.org/10.2478/jee-2024-0013
[2] UAVs for Disaster Management - An Exploratory Review - Julian Kedysa,b, Igor Tchappia, Amro
Najjarc (2023)
https://doi.org/10.1016/j.procs.2023.12.184
[3] Jianwen Xu, Kaoru Ota, and Mianxiong Dong, “Big Data on the Edge Computing for Disaster
Management” 7, Dec. 2020.
https://doi.org/10.4567/j.procs.2020.567
[4] K. G. Panda, S. Das, D. Sen, and W. Arif, “Design and Deployment of UAV Aided Post Disaster
Emergency Network” 2019.
[5] M. Deryck, J. Wyckmans, W. Joseph, and L. Martens, “Designing UAV aided emergency networks
for large-scale disaster scenarios” 2019

Dept of ECE, VKIT Page 20