sensors

Article
RDSP: Rapidly Deployable Wireless Ad Hoc System
for Post-Disaster Management
Ajmal Khan 1 , Adnan Munir 1 , Zeeshan Kaleem 2 , Farman Ullah 1 , Muhammad Bilal 3 ,
Lewis Nkenyereye 4 , Shahen Shah 1 , Long D. Nguyen 5 , S. M. Riazul Islam 6 and
Kyung-Sup Kwak 7, *
1 Department of Electrical and Computer Engineering, COMSATS University Islamabad-Attock Campus,
Attock 43600, Pakistan; [email protected] (A.K.); [email protected] (A.M.);
[email protected] (F.U.); [email protected] (S.S.)
2 Department of Electrical Engineering, COMSATS University Islamabad Wah Campus, Wah 47040, Pakistan;
[email protected]
3 Division of Computer and Electronics Systems Engineering, Hankuk University of Foreign Studies,
Yongin-si 17035, Korea; [email protected]
4 Department of Computer and Information Security, Sejong University, Seoul 05006, Korea;
[email protected]
5 Institute of Research and Development, Duy Tan University, Da Nang 550000, Vietnam;
[email protected]
6 Department of Computer Science and Engineering, Sejong University, Seoul 05006, Korea; [email protected]
7 Department of Information and Communication Engineering, Inha University, Incheon 22212, Korea
* Correspondence: [email protected]




Received: 9 November 2019; Accepted: 13 January 2020; Published: 19 January 2020


Abstract: In post-disaster scenarios, such as after floods, earthquakes, and in war zones, the cellular
communication infrastructure may be destroyed or seriously disrupted. In such emergency scenarios,
it becomes very important for first aid responders to communicate with other rescue teams in order
to provide feedback to both the central office and the disaster survivors. To address this issue, rapidly
deployable systems are required to re-establish connectivity and assist users and first responders
in the region of incident. In this work, we describe the design, implementation, and evaluation
of a rapidly deployable system for first response applications in post-disaster situations, named
RDSP. The proposed system helps early rescue responders and victims by sharing their location
information to remotely located servers by utilizing a novel routing scheme. This novel routing
scheme consists of the Dynamic ID Assignment (DIA) algorithm and the Minimum Maximum
Neighbor (MMN) algorithm. The DIA algorithm is used by relay devices to dynamically select
their IDs on the basis of all the available IDs of networks. Whereas, the MMN algorithm is used by
the client and relay devices to dynamically select their next neighbor relays for the transmission of
messages. The RDSP contains three devices; the client device sends the victim’s location information
to the server, the relay device relays information between client and server device, the server device
receives messages from the client device to alert the rescue team. We deployed and evaluated our
system in the outdoor environment of the university campus. The experimental results show that
the RDSP system reduces the message delivery delay and improves the message delivery ratio with
lower communication overhead.

Keywords: disaster management services; device-to-device communication; multi-hop relaying;

WiFi; ad hoc network; GPS

Sensors 2020, 20, 548; doi:10.3390/s20020548 www.mdpi.com/journal/sensors

Sensors 2020, 20, 548 2 of 21

1. Introduction
Cellular telephony is an extensively used communication technology. There are approximately
eight billion active cellular subscriptions globally with approximately half of those users added in
the last few years, mostly in developing areas [1]. Currently, mobile-cellular subscribers are more
than the total population of the world. This is because people enjoy more than one subscription to
take advantage of competing data plans of different cellular operators and so forth. Therefore, in
many developing areas, cellular networks have replaced conventional landline telephone systems
because of easy usage and the low cost of deployment. However, natural or man-made disasters
can disrupt or fully destroy the cellular and land line telecommunication infrastructures and services
in the effected areas. Recently, in September 2014, Hurricane Odile struck Mexico’s Baja California
coasts [2]. The hurricane of category four destroyed many towns, causing the massive destruction of
the electrical infrastructure that left over 90% of the population without electricity. The destruction of
the communication infrastructure resulted in the absence of cooperation between the aid organizations;
consequently, thousands of individuals endured hardship, due to mismanaged rescue operations.
After the distractions of a disaster, the communication among rescue groups, such as firemen, police
officers, and paramedics is vital. In particular, the feedback of first rescue responders is extremely
important for an effective rescue and restoration operation. From a networking perspective, the aim is
to recover connectivity to offer at least temporary communication services to rescue groups. One way
to overcome these issues is to arrange network components, such as relays, access points, or routers to
create a temporary network on request [3]. This needs a quickly deployable network to perform the
required relief efforts, including helicopters and first responders on the floor that can save many lives.
In addition, to guarantee the safety of survivors in the disaster-affected region, it is very important
to manage stranded people’s requests in a timely manner to provide a general picture of the total
injuries, relocation method, emergency needs, and so forth [4,5]. Furthermore, in an attempt to rapidly
handle and deliver food and other resources to displaced inhabitants, the need for a secure and reliable
communication network that is easy to deploy and relies on radio waves instead of a data cable would
be a good choice for communication in disaster situations [6].
In this work, we propose a rapidly deployable system using multi-hop relays for a post-disaster
wireless ad hoc network, named RDSP. The RDSP aims to reduce the average waiting times for
transmitting the rescue groups and victim’s location information towards the control server. In addition,
the RDSP scheme enables intermediate relay devices to dynamically select their IDs on the basis of
the information provided by their neighbor relays and then each intermediate relay selects the best
forwarders towards the control server to minimize end-to-end and round-trip message delivery delays.
Finally, the client device is used to transmit victim’s information via Wi-Fi towards the control server,
which then alerts emergency rescue teams to deliver food and other resources to displaced and
stuck survivors.
The remainder of this paper is organized as follows. Section 2 summarizes the related work.
Section 3 explains the flow charts and algorithms for the client, relay, and server devices. Section 4
describes the performance evaluation and compares the performance of the RDSP with the existing
scheme. Finally, Section 5 provides the concluding remarks.

2. Related Work
Improving the emergency response times in post-disaster situations, where the basic
communication infrastructure is completely dismantled, is a critical and challenging task. In such a
situation, rapidly deployable networks are needed to enable first responders to interact with disaster
survivors, each other, and the control room. These networks operate under challenging conditions,
such as power constraints and establishing a network back haul; further, the network must be easy to
deploy, operate, and maintain. To mobilize the smooth transit of rescue teams for providing emergency
services in disaster situations, a number of disaster management schemes or rapid deployment
Sensors 2020, 20, 548 3 of 21

systems have been proposed in the literature that mainly focus on emergency communication in order
to connect first responders to the control server.
Authors in [4,7,8] demonstrated the importance of rapidly deployable systems in disastrous
scenarios, and provided a survey of the number of schemes that offer adequate post-disaster
emergency services. They further discussed the features of the existing systems in the post-disaster
environment including their advantages and disadvantages. The authors in [7] also looked at the
networking parameters that are essential in a disaster environment, such as routing overhead, topology
management, energy efficiency, and multimedia bandwidth usage.
The authors in [5] have developed an energy efficient routing protocol that limits the number of
duplicate messages transmission to improve the data delivery ratio and extend the operating time of
battery-powered devices. The protocol reduces duplicate messages by finding recurring contacts and
generates a routing table that uses recurring contacts to transmit a message to a destination. By finding
repeated contacts and creating a routing table that utilizes these repeated contacts to send a message
to a destination, the proposed protocol considerably reduces the number of control and operation
management messages. Owing to reduction in message transmissions, their protocol improved the
overall network energy consumption, while maintaining a high delivery rate.
In [9], the authors have proposed an infrastructure independent device-to-device decentralized
network system. Various devices, such as GPS, camera, sensors, and transceivers communicate
without centralized coordination. The system works without any base station or access point, as
devices communicate with each other when the infrastructure is not available because of an accident
or emergency. For the communication within a local vicinity, the devices adopt TDMA to assign a
specific slot to each device for communication purpose.
In [10], the authors proposed the concept of rescue base stations (RBS). The RBS is a
GSM-compatible solar power drop-in communication system especially designed for disaster scenarios.
The proposed system consists of a number of disconnected RBS(es) that provide GSM facilities to a
number of registered individuals lying in their coverage range. Each RBS locally stores call/SMS data.
Since RBS(es) are disconnected and do not share any network link, individuals with android devices
act as information carriers between disconnected RBS(es) to transport the necessary information from
one RBS to another.
In [11], the authors proposed a system to help military officers in critical situations, such as
war conditions or attacks in a gangster area. In military warfare, a robust communication system is
required so that the military head can communicate with the soldiers and relay the information easily.
A mobile robot is utilized to carry and deploy the nodes at the scene of incident. However, nodes have
restricted ranges and can be damaged as the robot moves around snags. The proposed system has
communication limitations in situations where no line of sight will be available, such as in urban area.
In [12], the authors proposed a robot-assisted scheme, which assists the intermediate nodes
between source and destination to relay a message over a long distance. The proposed system deploys
robots that utilize mesh technology to create autonomous broadband wireless networks. The actions
of robots are controlled by relative signal strength indicators (RSSI). By redistributing the network
nodes, it is possible to increase the existing system’s throughput. The system is adaptable, self-forming,
and self-healing.
Energy in rapidly deployable networks is a major constraint, because the intermediate relays
consume a high amount of energy while establishing a connection with other nodes, exchanging
information, and routing critical information towards the command center [13,14]. Since wireless
relays are battery powered and have a limited power supply, an energy-efficient routing protocol is
essential for rapidly deployable networks.
The authors in [15] outlined an ad hoc airborne communication scheme using balloons, having
excellent line-of-sight, wide transmission range, and low interference. The flying balloons create a
multi-hop ad hoc network and get access to internet through an internet gateway placed in disaster hit
Sensors 2020, 20, 548 4 of 21

areas. The rescue workers access internet services by connecting the flying balloons. On each balloon,
a Global Positioning System (GPS) receiver is installed to find the balloon’s position.
In work [16], the authors proposed a Movable and Deployable Resource Unit (MDRU). The goal
of the MDRU is to transport resources to a disaster location and set up the network to provide the
necessary communication services after a disaster. A van-type MDRU was deployed to establish a
temporary network in disaster zones. The van carries all necessary equipment required to establish
ad hoc communication at the disaster area. The authors in [17] suggested a model to enhance the
MDRU-based network’s energy resource utilization and spectrum improvement. The proposed model
consists of two stages, named topology formation and transmission division. The former stage
configures the gateways of the k spectrum and the later stage divides the transmission from the sender
gateways to the MDRU resource unit.
In [18–25], the authors proposed various deployment schemes for disaster response to a building
collapse, search and rescue, as well as a resource request through ad hoc networks. In [18], the
authors proposed a novel approach, called Supporting Urban Preparedness and Emergency Response
using Mobile Ad hoc Network (SUPER-MAN). The main objective of that approach was to enable
structural engineers and first responders to efficiently disseminate a damaged building status and
resource request information towards the control sever with minimum possible interference and
delays. The proposed SUPER-MAN system relies on Radio Frequency Identification (RFID) tags to
store assessment information during disaster. A Mobile Ad hoc Network (MANET) of RFID tags is
established where Dynamic Source Routing (DSR) is implemented as the routing protocol.
In [19], the authors proposed a novel method using a ground penetrating radar (GPR) that
automatically detects, locates, and characterizes empty space within disaster rubble. In the proposed
method, radargrams are preprocessed to segment the boundaries of empty spaces on the basis of
radar signal patterns. To search for uncertainties, 95% confidence intervals are constructed around
the segmented boundaries. The geometric relations of the detected boundaries and their signal
characteristics are examined to confirm the existence of free space and to improve detection accuracy.
Then, the calibrated velocity of a radar wave and its travel time are used to estimate the location and
dimension of empty spaces or voids.
In [20], the authors proposed an emergency resource repository portal (E2RP) system, which is a
web-based geo-database service that enables on-site and off-site decision makers to access resource
information. The whole E2RP framework incorporates a web collaboration service, radio frequency
identification RFID tags, a building blackbox system BBS, a geo-database, and a geographic information
system GIS. The E2RP framework provides first responders, including civil engineers, a collaboration
medium that enables them to actively respond to disasters. The framework also provides access to
critical building information through the BBS. RFID tags are used to store building information which
is accessible to first responders through the wireless adhoc network. The GIS is used to locate, collect,
and distribute resources to first responders.
In [21], a Geographic Information System (GIS)-based framework is proposed that facilitates
equipment allocation during disasters. The proposed framework incorporates three subsystems to
assist in information gathering and decision making. First, an application is developed that runs
on mobile devices to request on-field resources. Second, a resource repository is deployed with a
geospatial database that allows a graphical interface to spatially query resources. Additionally, a GIS
is introduced that allows for automatic decision-making, such as matching resources and identifying
routes for resource distribution. The proposed framework incorporates decision models into the
system to assist complex decision-making during equipment delivery.
In [22], the authors proposed a state-of-the-art technique for collecting data and extracting
information to avoid disaster-related injury and post-event damage. A database repository based
on GIS, called Extreme Events Database Viewer (EEWV), is being developed to store spatial and
temporal data that defines communities before and after disasters. This web platform can store
multiple geolocated data formats including photographs and 360° videos. A tool was designed to
Sensors 2020, 20, 548 5 of 21

automatically extract photographs from 360° video data. Extracted images provide a manageable
data set to efficiently document the characteristics of buildings and the surrounding environment.
The propose system’s main objective was to find buildings that were vulnerable to floods and storms.
To this end, 1950 buildings were filmed passively with a 360° camera mounted on the vehicle. In order
to train a deep learning neural network, these extracted building images were used by the neural
network to determine whether a building was elevated or not.
In [23], the authors proposed a Reliable Routing Technique (RRT) that ensures reliable data
delivery towards the destination device, using mobile devices that are carried by moving people in
the incident area. Each mobile device broadcasts the received message towards the destination by
maintaining a priority list of probable forwarding candidates. The proposed RRT technique guarantees
that the second priority candidate will forward the data packet to the destination device if the first
priority candidate is unable to forward the data packet due to its mobility, thus ensuring the reliability
of data delivery in the network.
The concept of breadcrumbs is introduced by the authors in [26–29]. Breadcrumbs are tiny and
cost-efficient relay devices. Their only objective is to relay packets between edge nodes. Thus, in
disaster scenarios, rescue team members carry several breadcrumb devices along with a mobile radio
to communicate with command center via breadcrumb devices. Rescue team members must regularly
drop breadcrumb devices as they explore the disaster area in order to retain end-to-end connectivity
with the control server. The breadcrumb relays are dropped to create a static ad hoc network on demand.
The command center retains contact with the rescue teams members via relays dropped by the rescue
team to enlarge the coverage area. The breadcrumb approach guarantees reliable communication,
offers an increased coverage area and eliminates the probability of network partitioning.
Extensive research was carried out to address the problem of the decision of the deployment of
breadcrumb. Each proposal describes its own deployment algorithms but they have various common
features. Some algorithms monitor the link quality by measuring the signal-to-noise ratio (SNR) [26],
bandwidth [27], or received signal strength indicator (RSSI) [28]. A threshold is set to trigger a
deployment event. When the quality of the link falls below this limit, a fresh relay must be dropped
by the user. For instance, a pre-defined threshold is used for all applications [26] and [28], whereas
in [27], the threshold is set based on the bandwidth requirement of each application. In [28], the
authors have proposed an ad hoc deployment system to efficiently communicate with victims during
disasters. In the proposed system, two types of control information are used for the deployment of
relay i.e., relay is deployed either because the link quality degradation is detected by the mobile user,
or because an explicit relay deployment request is received by the mobile user from its neighbors.
The disaster area could be sufficiently large; therefore, relays support multiple hop transmission to
leverage end-to-end communication among each other. The command center is located on the outskirts
of the incident region and members of the rescue team begin moving from the command center and
move in separate directions into the incident scene. It is assumed that mobile users are connected
with each other and with the command center at any time. A mobile rescue team member can drop a
relay device if required to maintain connectivity. Each mobile user determines where and when a relay
device should be deployed by running an algorithm and alerting the host user through some devices,
such as blinking or strong light or sound, when it is necessary to unfold the network. Similarly, in [26],
the authors proposed an algorithm where relays perform a rapid evaluation of a physical layer to
decide on the deployment of the next relay. The relay constantly transmits probe packets to the relays
that have been dropped before. A receiving relay responds with an acknowledgment packet if it is
within the communication range. Then, through ACK reception, the transmitting relay calculates the
SNR; if the value of the SNR drops below the threshold level, a new relay is dropped.
The breadcrumb approach provides a suitable solution to extend the coverage area for rescue
teams in disaster situations. However, this approach does not offer redeployment possibilities because
the relays do not have their own mobility. Indeed, as mentioned in the above proposals, the first
responders must drop breadcrumb devices to set up an ad hoc network. However, this is not necessarily
Sensors 2020, 20, 548 6 of 21

the perfect solution. This is because when the first responders join the rescue operation, relay
deployment is not their first priority. Hence, they may forget to drop a relay or merely miss the
deployment signal. To solve this problem, an automatic breadcrumb dispenser is proposed in [29].
A Utility Function (UF)-based algorithm is proposed that sets criteria to deploy new breadcrumbs
automatically. The UF based algorithm works as follows: the requester broadcasts a help message to
initiate the algorithm. After receiving a help message, all the neighbors of requester send their data
(number of breadcrumbs) to the requester. Following a predefined timeout, the requester calculates
the value of each of its neighbors’ utility functions and transmits a drop message to a neighbor with
the highest UF value to deploy a new breadcrumb.
As described above, the existing breadcrumb approaches only focus on the efficient deployment
of breadcrumbs to enlarge the coverage area and eliminate the probability of network partitioning.
These approaches, however, do not provide effective routing schemes to deliver emergency request
messages with minimum latency to the command center. Indeed, these breadcrumb approaches utilize
existing routing protocols already developed for mobile ad hoc networks. This is not, however, a
perfect solution. This is because breadcrumb devices are utilized as battery-driven intermediate relays,
and the current routing schemes can rapidly deplete the battery life of these devices. To efficiently
utilize the battery life of breadcrumb devices, the network protocol should be designed in a manner to
create the shortest multi-hop path between the command server and the rescue team with a reduced
control message overhead and a minimum end-to-end message delivery delay. Furthermore, the
current breadcrumb approaches do not provide the command server with any data regarding the
location of first responders, which ultimately makes it very difficult to find the victims and rescue team
positions in post-disaster situations. To the best of our knowledge, no work exists that both designs a
unique routing protocol for breadcrumb devices and incorporates location information of victims to
determine their distance from the command server.
Therefore, the major contribution of this study comprises a rapidly deployable system that both
delivers request messages to the command server by utilizing novel routing schemes and also manages
the location information of victims to calculate their distance from the command server. The novel
routing scheme consists of a Dynamic ID Assignment (DIA) algorithm and a Minimum Maximum
Neighbor (MMN) algorithm. The DIA algorithm is used by relay devices to dynamically select their
IDs on the basis of all available IDs of networks. Whereas, the MMN algorithm is used by the client
and relay devices to dynamically select their next neighbor relays for the transmission of messages.
In addition, we provide details of algorithms performed by the client device, relay device, and server
device. Furthermore, extensive real time experiments are preformed to demonstrate how the proposed
RDSP scheme reduces the control messages’ overhead to deliver the request messages with a minimum
end-to-end delay and an increased message delivery ratio in post disaster situations.

3. System Architecture of the RDSP Scheme

This section presents the architecture of the RDSP scheme, which aims to reduce the average
waiting times for transmitting victim’s information towards the server by utilizing the following
key features:

• Client Device: The client devices are the end point communication devices held by rescue team
members and victims. The client device is a WiFi-enabled device that manages to transmit the
victim’s location information to the server device using relay devices. The client devices are used
to establish the communication between rescue teams and server and to send feedback about the
emergency situation.
• Relay Device: Relay devices are randomly deployed at a distance of 90 m from each other to
transmit the victim’s location information generated by the client device towards the server and
then send back acknowledgment information to the client device. Relay devices dynamically
connect with each other in a manner to establish the shortest path towards the server.
Sensors 2020, 20, 548 7 of 21

• Server Device: The server device continuously listens for the arrival of incoming messages sent by
the client device via relay devices in order to alert the rescue teams. Moreover, it sends the control
and operational messages and is also responsible for receiving the feedback from rescue teams.

Figure 1 illustrates a deployment scenario of the RDSP system after disaster. When disaster occurs,
the rescue team members will deploy the RDSP system as follows: A server device is installed in
the incident area that will receive updates from rescue teams and victims. Additionally, rescue team
members will move forward towards disaster areas while deploying relay devices at equal distances
of about 90 m until any victim is sighted. Then, the client device that is carried by the rescue team
members is used to send the victim’s position information to the server. The distance of 90 m between
relays is managed by incorporating GPS modules in relays. After the server module is installed in an
incident area, we start deploying the relays as follows: First of all, the first relay wirelessly connects
with the server module. Afterwards, while carrying the relay and moving away from the server
module, the relay continuously calculates its distance from the server using the GPS module that
provides location coordinates of the relay module, whereas the location coordinates of the server are
fixed and known to the relay module. If the calculated distance is 90 m, a green LED (installed on
relay module) lights up, indicating to drop the first relay module at that particular position. Similarly,
a second relay is chosen which connects with the first relay and the distance between the first relay
and the second relay is again calculated by the second relay and it is dropped when the green LED
on the second relay lights up. Following this procedure, all the relay devices manage to maintain a
distance of 90 m between each other. The 90 m distance was selected because the relay device uses WiFi
technology that has a transmission range of 90 m. However, the transmission range can be increased by
using other advanced technologies, which is highly application dependent. It is indicated in Figure 1
that both the client device and relay device communicate wirelessly with the server device via WiFi.
The server device manages all the request messages received and sends back an acknowledgment to
the client device to ensure the reception of the request message.

Figure 1. Deployment scenario of the RDSP scheme.

Figure 2 presents a flow chart of the server device. After the initialization, the server waits for the
arrival of request messages. If a message arrives, the server sends back an acknowledgment message
to the client device via relay devices and generates an alarm message to alert the rescue teams.
Sensors 2020, 20, 548 8 of 21

Figure 2. Flow chart of server device.

Algorithm 1 presents the operation of the microcontroller in the server device. The input to the
server device is a request message Rmsg , whereas the output is an acknowledgment message and an
alarm message. As shown in step 2, the server device continuously waits for the detection of Rmsg .
If Rmsg is detected, alarm message is generated, acknowledgment is sent back to the client device and
Rmsg is printed on screen. The microcontroller in the server device utilizes the Haversine formula [30]
to determine the distance between the client device and server. This process is defined by the following
Equation (1):
s
∆φ ∆δ
   
−1 2 2
dist = 2rsin sin + cos (φ1 ) * cos (φ2 ) *sin (1)
2 2

where φ is the latitude, δ is longitude (in radians).

Algorithm 1 Server device

Input:Rmsg (Request message containing node ID, message ID, and GPS coordinates)
Output: Ack msg (Acknowledgment message )
Alarmmsg (Alarm message generated)

1: procedure
2: Step 1: defining and initializing variables
3: Step 2: detecting events
4: while 1 do
5: if Rmsg = true then
6: Generate Alarmmsg
7: send Ack msg back
8: Print request message on screen.
9: end if
10: end while
11: end procedure
Sensors 2020, 20, 548 9 of 21

Figure 3 shows the flow chart of the relay device. After initialization, the relay device scans
for available networks and applies the Dynamic ID Assignment (DIA) algorithm (explained later in
Algorithm 2) to generate its own ID. Then, the Minimum Maximum Neighbor (MMN) algorithm is
utilized that returns minimum and maximum IDs (explained later). Afterwards, the relay device waits
for the arrival of messages. If the request message arrives, it selects the minimum ID to transmit the
message to the server. The relay with the minimum ID is selected because it is much closer to the
server (as will be explained later in the DIA algorithm) and delivers the message in minimum possible
time. However, if an Ack message arrives, the relay device selects the maximum ID to transmit the
message back to the client.

Figure 3. Flow chart of relay device.

Algorithm 2 Dynamic ID Assignment (DIA)

Input: AVBnet (Available Networks)
Output: IDassigned (Assigned ID )

1: procedure
2: Step 1: defining and initializing variables
3: Arrayid = Array containing available networks IDs
4: Findex = First index of Arrayid
5: Sindex = Second index of Arrayid
6: Step 2: Assigning ID to relay
7: Store AVBnet in Arrayid
8: Sor tArrayid in ascending order
9: IDassigned = Arrayid [ Findex ]
10: if IDassigned = −1 then
11: IDassigned = Arrayid [ Sindex ]
12: end if
13: IDassigned = IDassigned +1
14: end procedure

Algorithm 2 presents the dynamic ID assignment (DIA) algorithm. This DIA algorithm is used by
relay devices to dynamically select their IDs based on all the available IDs of networks. Initially, it is
assumed that the server has an ID = 0 and all the deployed relay devices have an ID = −1. As shown in
step 2, each relay device stores available network IDs in Arrayid . Then, Arrayid is sorted in ascending
order. Finally, the relay device generates its own ID by selecting the first index of Arrayid as it the
contains minimum ID. However, if the first index of Arrayid contains −1, then the second index of
Arrayid is selected and incremented by 1.
Sensors 2020, 20, 548 10 of 21

The detailed procedure of the DIA algorithm is explained in Figure 4. All relays are deployed
randomly at a distance of 90m and initially their IDs are −1 and the server ID are 0, as shown in
Figure 4a. According to Figure 4b, relay X will receive network IDs 0 and −1 from server S and relay Y,
respectively. By applying the DIA algorithm, relay X will choose the positive minimum ID, i.e., 0 and
increments it by 1. Therefore, the ID of relay X becomes 1. Similarly, in the same fashion, relay Y will
receive IDs 1 and −1 from relay X and Z, respectively. As shown in Figure 4c, after applying the DIA
algorithm, relay Y will chose the positive minimum ID, i.e., 1 and increments it by 1. Therefore, the ID
of relay Y will become 2 and this process will continue until all the relays will be assigned dynamic
IDs, as shown in Figure 4e. Hence, it can be seen from Figure 4 that relays having lower IDs are much
closer to the server as compared with relays having higher IDs.

Figure 4. Dynamic ID assignment procedure.

Algorithm 3 presents the MMN algorithm. This MMN algorithm is used by client and relay
devices to dynamically select their next neighbor relays for the transmission of messages. As shown in
step 2, each relay device stores available network IDs in Arrayid . Then, Arrayid is sorted in ascending
order. Finally, the relay device finds Min ID and Max ID by selecting the first and last index of Arrayid ,
respectively.

Algorithm 3 Minimum Maximum Neighbor (MMN)

Input: AVBnet (Available Networks)
Output: Min ID and Max ID

1: procedure
2: Step 1: defining and initializing variables
3: Arrayid = Array containing available networks IDs
4: Findex = First index of Arrayid
5: Lindex = Last index of Arrayid
6: Step 2: Finding Minimum and Maximum IDs
7: Store AVBnet in Arrayid
8: Sort Arrayid in ascending order
9: Min ID = Arrayid [ Findex ]
10: Max ID = Arrayid [ Lindex ]
11: end procedure
Sensors 2020, 20, 548 11 of 21

The detailed procedure of the MMN algorithm is explained in Figure 5. Figure 5a shows that all
the relays have been assigned IDs after applying the DIA algorithm, as explained earlier in Algorithm 2.
Afterwards, each relay will select its next neighbor relay for the transmission of messages. It is shown
in Figure 5b that relay X will receive IDs 0 and 2 from server S and relay Y respectively. By applying
MMN algorithm, relay A will choose 0 as the minimum ID and 2 as maximum ID. Similarly, in the
same fashion, relay Y will receive IDs 1 and 3 from relay X and Z, respectively. After applying the
MMN algorithm, relay Y will chose 1 as the minimum ID and 3 as the maximum ID.

Figure 5. Minimum Maximum Neighbor selection

.

Algorithm 4 presents the code for the microcontroller in the relay device. The input to the relay
device is a request message Rmsg , an acknowledgment message Ack msg , and available networks AVBnet .
The output is either a request message or an acknowledgment message. As shown in step 1, the relay
device scans for available networks and then it applies the DIA algorithm to select its ID and then
the MMN algorithm to select minimum and maximum neighbor IDs. Then in step 2, the relay device
continuously waits for the detection of events. If Rmsg is detected, then the request message is sent to
the next relay node having Min ID . However, if Ack msg is detected, then acknowledgment is sent back
towards the client device via the next relay node having Max ID .

Algorithm 4 Relay device

Input:Rmsg (Request message containing node ID, message ID and GPS coordinates)
Ack msg (Acknowledgment message )
AVBnet (Available Networks)
Output:Rmsg (Request message containing node ID, message ID and GPS coordinates)
Ack msg (Acknowledgment message )

1: procedure
2: Step 1: defining and initializing variables
3: Scan AVBnet
4: Apply DIA Algorithm (assigns ID)
5: Apply MMN Algorithm ( returns Min & Max ID)
6: Step 2: detecting events
7: while 1 do
8: if Rmsg = true then
9: send Rmsg to Min ID
10: end if
11: if Ack msg = true then
12: send Ack msg to Max ID
13: end if
14: end while
15: end procedure
Sensors 2020, 20, 548 12 of 21

Figure 6 shows the flow chart of client device. The client device includes a GPS system,
a microcontroller, WiFi device, and a push button. The GPS device is utilized to receive the latitude and
longitude information of the victim. The microcontroller is utilized to transmit the victim’s position
information to the server via the WiFi module. As shown in the flowchart, after initialization, the
client device scans for available networks and applies the MMN algorithm that returns the minimum
and maximum IDs (explained earlier). Afterwards, the client device reads the status of push button.
If the push button is pressed, it connects with the relay node having a minimum ID. The client device
then reads the GPS coordinates and transmits the request message towards the server via multi hop
intermediate relays. The relay with a minimum ID is selected because it is much closer to the server
(as explained earlier in DIA algorithm). After sending the request message, the client device waits
for the arrival of the acknowledgment message. If acknowledgment is received, it is then printed on
the screen.

Figure 6. Client device.

Algorithm 5 presents the code for the microcontroller in the client device. The input to the client
device is the push button Pb , acknowledgment message Ack msg , GPS coordinates GPScor , and available
networks AVBnet . The output is a request message Rmsg . As shown in step 1, the client device scans
for available networks and it applies the MMN algorithm to select minimum and maximum IDs. Then,
in step 2, the client device continuously waits for the detection of events. Since, the client device
includes a push button that is pressed by a rescue team member if any victim is sighted. Therefore, if
the push button Pb is detected, the client device reads GPScor and transmits the request message to the
next relay node having Min ID and continuously waits for the arrival of the acknowledgment message.
If the Ack msg is detected, then acknowledgment is printed on the screen.
Figure 7a shows a sample network scenario to explain the working procedure of the RDSP system.
As shown in the figure, relays are deployed in a manner to establish two different paths between server
device S and client device C. The first path is defined by C, a1 , a2 , a3 , S whereas the second path is
defined by C, a10 , a9 , a8 , a7 , a6 , a5 , a4 , S. After deployment, each relay node applies the DIA algorithm
to select the dynamic ID based on available networks.
As shown in Figure 7a, after applying the DIA algorithm, the IDs of a3 , a2 , a1 are 1, 2, and 3
respectively. Similarly IDs of a4 , a5 , a6 , a7 , a8 , a9 , a10 are 1, 2, 3, 4, 5, 6, 7, and 8 respectively. After ID
assignments, each relay and client node applies the MMN algorithm to to select Min ID and Max ID .
Client C receives ID 3 from relay a1 and ID 8 from relay a10 . Therefore, client C selects 3 as the
minimum ID and 8 as the maximum ID. Similarly, relay a1 selects 2 as the minimum ID and relay a10
selects 7 as the minimum ID. Both relay a1 and relay a10 only receive IDs from relay a2 and relay a9
respectively but not from client C as client C has no ID. Relay a2 receives ID 1 and 3 from relays a3
Sensors 2020, 20, 548 13 of 21

and a1 , respectively, and hence, selects 1 as the minimum and 3 as the maximum ID. This process is
continued until all relays select their minimum and maximum neighbor IDs.

Algorithm 5 Client device

Input:Pb (Push button)
Ack msg (Acknowledgment message )
GPScor (GPS coordinates )
AVBnet (Available Networks)
Output:Rmsg (Request message containing node ID, message ID, and GPS coordinates)

1: procedure
2: Step 1: defining and initializing variables
3: Scan AVBnet
4: Apply MMN Algorithm ( returns Min & Max ID)
5: Step 2: detecting events
6: while 1 do
7: if Pb = true then
8: read GPScor
9: send Rmsg to Min ID
10: end if
11: if Ack msg = true then
12: Pring Ack msg on screen
13: end if
14: end while
15: end procedure

Figure 7b shows how the request message Rmsg is sent from client C to server S. According to
Algorithm 5, if the push button is pressed, the client device sends the request message to a neighbor
having Min ID , that is, relay a1 having ID 3. Therefore, client C sends the request message to relay a1
instead of relay a10 . According to Algorithm 4, if the request message is received, the relay device sends
it to a neighbor having Min ID . Therefore, relay a1 sends the request message to relay a2 . Afterwards,
relay a2 sends it to relay a3 having ID 1. Finally, relay a3 sends the request message to server S having
ID 0.
Figure 7c shows how the Ack message is sent back to the client device from server S. According to
Algorithm 1, server S sends back the acknowledgment message to relay a3 . As described in Algorithm 4,
if the acknowledgment message is received, the relay device sends the received acknowledgment
message to a neighbor having Max ID . Therefore, relay a3 sends the acknowledgment message to
relay a2 having ID 2. Afterwards, relay a2 sends it to relay a1 having ID 3. Finally, relay a1 sends it to
client C.

Figure 7. Cont.
Sensors 2020, 20, 548 14 of 21

Figure 7. Sample network deployment scenario.

4. Performance Evaluation
In this section, the performance of the RDSP scheme is compared with the UF scheme [29] by
deploying the real time multihop communication network.

4.1. Deployment Environment

A multihop communication network is deployed in Comsats university Islamabad-attock campus
as shown in Figure 8.
The total area of the university is 153,000 m². As shown in the figure, the server device is installed
at point S and the client device is installed at point C. The relays are deployed all over the area in a
manner to establish four different paths between server device S and client device C. The first path
defined by points C, Y, Z, S contains 7 relays, second path defined by points C, Z, S contains 6 relays,
third path defined by points C, S contains 5 relays and fourth path defined by points C, X, S contains 8
relays. The relays are placed at fixed positions at equal distance of 90 m. To find the shortest paths
between the client and server devices, the proposed RDSP system utilizes DIA and MMN algorithms
whereas the UF scheme uses Destination-Sequenced Distance Vector (DSDV) [28] routing algorithm.
In the DSDV, each node maintains a routing table with routes to destinations. An entry for a given
destination consists of the ID of the next-hop, total hops to the destination, route’s sequence number,
and the time for the recent route update. In the case of the proposed RDSP, each node periodically
broadcasts a hello message every 2 s containing only the node ID. However, in the UF scheme, each
node periodically broadcasts an entire routing table every 1 s containing the route destination, the
advertising node’s next-hop node for that route, the number of hops to the destination, and the
sequence number. Following the deployment of the multihop communication system, both the client
device and server device are tested to communicate wirelessly with each other across all four paths.
Sensors 2020, 20, 548 15 of 21

One hundred request messages were generated to collect the results for each path. The experiment
duration was set at 550 s. The results presented for each path are averaged over 5 repeated experiments.
The experimental parameters are summarized in Table 1.

Figure 8. Deployment of the RDSP system.

Table 1. Experimental Parameters

Parameter Value
Deployment scenario COMSATS University Campus
Deployment area 153,000 m²
Total relays 23
Total Paths 4
Relays Transmission range 90 m
Routing protocol (RDSP) DIA and MMN
Routing protocol (UF) DSDV
Hello interval (RDSP) 2 sec
Hello inerval (UF) 1 sec
Request messages 100
Experiment duration 550 s

4.2. Performance Metrics

To investigate the performance of the RDSP scheme, the following metrics were used:

• Distance covered: refers to the total distance covered by intermediate relays using wireless
transmission
• End-to-end delay: refers to the average time it takes for a request message sent from the client
device to reach the server device.
• Round-trip delay: refers to the average time it takes for a message sent from the client device to
reach the server device and an acknowledgment message sent back from the server device to
reach the client device.
• Message delivery ratio: refers to the percentage of messages successfully received by the
server device.
• Network overhead: refers to the total number of control messages sent by relays on different paths.
Sensors 2020, 20, 548 16 of 21

4.3. Distance Covered

Figure 9a shows the distance covered by different edges described in Figure 8 along with the
deployed number of relays at each edge. Edge C-Y covers 358 m with 4 relays whereas edge Y-Z and
Z-S cover 90 and 268 m with 2 and 3 relays respectively. Similarly, edge C-Z and C-S cover 352 and
528 m with 4 and 5 relays, respectively. Finally, edge C-X and X-S cover 352 and 441 m with 4 and
5 relays, respectively. It is evident from Figure 9a that as the number of relays increases across an
edge, the distance covered also increases and vice versa. Figure 9b shows the distance covered by four
different paths described in Figure 8 along with the number of relays deployed on each path. Distance
covered by path-1 is 716 m and comprises of three edges, i.e., C-Y, Y-Z and Z-S. Similarly, the distance
covered by path-2 is 628 m and comprises two edges, i.e., C-Z and Z-S. Likewise, path-3 covers 582
m and comprises of edge C-S whereas path-4 covers 793 m and comprises of edges C-X and X-S. It
is once again obvious from Figure 9b that the paths comprising of more relays cover long distances
compared with the paths comprising of less relays.

Figure 9. Distance covered by edges and paths.

4.4. End-to-End Delay

Figure 10a compares the end-to-end delay for the RDSP system and UF scheme. It shows that
the end-to-end delay for both systems increases as the number of relays increases across the path.
To elaborate, the end-to-end delay for path-4 having 8 relays is greater than that of path-1 having 7
relays. Similarly, the end-to-end delay of path-1 is greater than that of path-2 having 6 relays. Likewise,
Sensors 2020, 20, 548 17 of 21

the end-to-end delay of path-2 is greater than that of path-3 having 5 relays. This is because when the
number of relays increases across a particular path, the processing and transmission time of messages
also increase, but more relays offer a long-distance coverage benefit. However, the RDSP system
achieved around 12% lower end-to-end delays than the UF scheme across all the four paths. This
was because the UF scheme is based on a DSDV protocol that requires all the relays to periodically
exchange hello messages and entire routing tables, which leads to frequent contention and collisions
among neighboring relays. In such cases, the relays must wait for a busy channel to become idle
before performing any transmission. On the other hand, the RDSP system avoids the formation and
exchange of routing tables and creates routes between the server and client device on the fly during
the deployment of relays, as explained earlier in Figure 7, which eventually reduces collisions among
neighboring relays and causes a decrease in the end-to-end delay. Figure 10b compares the round-trip
delay for the RDSP system and UF scheme. It shows that the round-trip delays are approximately twice
the end-to-end delays of both systems and R the DSP system achieves around 13% lower round-trip
delays than the UF scheme across all the four paths.

Figure 10. End-to-end and round-trip delays.

4.5. Message Delivery Ratio

Figure 11 compares the message delivery ratio for the RDSP system and UF scheme. It shows
that the message delivery ration for both systems decreases as the number of relays increased across
the path. To elaborate, the packet delivery ratio for path-4 having 8 relays is less than that of path-1
having 7 relays. Similarly, the packet delivery ration of path-1 is less than that of path-2 having 6
relays. Likewise, the packet delivery ratio of path-2 is less than that of path-3 having 5 relays. This
Sensors 2020, 20, 548 18 of 21

was because when the number of relays increases across a particular path, the frequent contention
and collisions of messages among neighboring relays also increases which eventually reduces packet
delivery ratio. However, RDSP system achieved around 8% higher message delivery ratio than UF
scheme across all the four paths particularly at high relay densities because it avoids exchange of
routing tables thus reduces collisions among neighboring relays and hence achieves higher message
delivery ratio. On the other hand, UF scheme utilizes periodic exchange of entire routing tables
that leads to frequent contention and collisions among neighboring relays which eventually reduces
message delivery ratio.

Figure 11. Message delivery ratio.

4.6. Network Overhead

Figure 12 compares the average network overhead for the RDSP system and UF scheme. The UF
overhead was very high because of the periodic exchange of entire routing tables to maintain the
whole network information at each relay node. In contrast, the RDSP system overhead was much
lower, due to the absence of the exchange of routing tables. Thus, the overhead in the RDSP system
was related to the transmission of hello messages and acknowledgment messages. As a result, the
average network overhead for the RDSP was about 33% lower than that for the UF scheme.

Figure 12. Network overhead.

Sensors 2020, 20, 548 19 of 21

5. Conclusions
In this study, an RDSP was proposed that aims to reduce the average waiting times for transmitting
the request messages containing rescue groups and victim’s location information towards the control
server. Additionally, unlike existing schemes, the RDSP does not rely on the periodic exchange of entire
routing tables. However, the proposed RDSP scheme enables intermediate relays to dynamically select
their IDs based on the information provided by their neighbor relays and then each intermediate relay
selects the best forwarders towards the control server to minimize end-to-end delays. The results of
real time experiments demonstrate that with the proposed RDSP scheme, a request message generated
by the client device can reach the server device with a minimal delay in both light and heavily
deployed paths compared to the existing UF system. In addition, the results confirmed that the RDSP
outperforms the UF scheme under various relay densities in terms of the end-to-end delay, round-trip
delay, massage delivery ratio, and network overhead.
Author Contributions: conceptualization, hardware, software, writing, A.K. and A.M.; methodology, validation,
Z.K. and F.U.; resources, conceptualization, funding acquisition, M.B. and L.N.; writing–original draft preparation,
S.S.; resources, writing-review, and editing, L.D.N.; writing-review and editing, funding acquisition, S.M.R.I.
and K.-S.K. All authors have read and agreed to the published version of the manuscript.
Funding: This work was supported by National Research Foundation of Korea-Grant funded by the Korean
Government (Ministry of Science and ICT) NRF 2020R1A2B5B0200247
Conflicts of Interest: The authors declare no conflict of interest.

References
1. ITU. Measuring the Information Society Report 2018. Available online: https://www.itu.int/en/ITU-D/
Statistics/Pages/publications/misr2018.aspx (accessed on 2 June 2020).
2. Cangialosi, J.P.; Kimberlain, T.B. Tropical Cyclone Report; National Hurricane Center: Miami, FL, USA,
September 2014.
3. Mase, K. How to Deliver Your Message from/to a Disaster Area. IEEE Commun. Mag. 2011, 49, 52–57.
[CrossRef]
4. Miranda, K.; Molinaro, A.; Razafindralambo, T. A survey on rapidly deployable solutions for post-disaster
networks. IEEE Commun. Mag. 2016, 54, 117–123. [CrossRef]
5. Uddin, M.; Ahmadi, H.; Abdelzaher, T.; Kravets, R. A low-energy, multi-copy inter-contact routing protocol
for disaster response networks. In Proceedings of the 6th Annual IEEE Communications Society Conference
on Sensor, Mesh and Ad Hoc Communications and Networks, Rome, Italy, 22–26 June 2009; pp. 1–9.
6. Guo, W.; Huang, X. Mobility model and relay management for disaster area wireless networks.
In Proceedings of the Third International Conference (WASA 2008), Dallas, TX, USA, 26–28 October 2008;
pp. 274–285.
7. Jahir, Y.; Atiquzzaman, M.; Refai, H.; Paranjothi, A.; LoPresti, P. G. Routing protocols and architecture for
disaster area network: A survey. Ad Hoc Netw. 2019, 82, 1–14. [CrossRef]
8. Reina, D.; Askalani, M.; Toral, S.; Barrero, F.; Asimakopoulou, E.; Bessis, N. A Survey on Multihop Ad Hoc
Networks for Disaster Response Scenarios. Int. J. Distrib. Sens. Netw. 2015, 11, 647037. [CrossRef]
9. Li, H.; Shan, L.; Matsuda, T.; Miura, R. Design and deployment of infrastructure-independent D2D
networks without centralized coordination. In Proceedings of the 2015 International Symposium on Wireless
Communication Systems (ISWCS), Brussels, Belgium, 25–28 August 2015; pp. 376–380.
10. Ghaznavi, I.; Heimerl, K.; Muneer, U.; Hamid, A.; Ali, K.; Parikh, T.S.; Saif, U. Rescue base station.
In Proceedings of the Fifth ACM Symposium on Computing for Development, San Jose, CA, USA, 5–6
December 2014; doi:10.1155/2012/614532 [CrossRef]
11. Pezeshkian, N.; Nguyen, H.G.; Burmeister, A. Unmanned Ground Vehicle Radio Relay Deployment System
for Non-Line-of-Sight Operations. In Proceedings of 13th IASTED International Conference on Robotics and
Applications, Würzburg, Germany, 29–31 August 2007; pp. 501–506; doi: 10.21236/ada475525 [CrossRef]
12. Nguyen, C.Q.; Min, B.-C.; Matson, E.T.; Smith, A.H.; Dietz, J.E.; Kim, D. Using mobile robots to establish
mobile wireless mesh networks and increase network throughput. Int. J. Distrib. Sens. Netw. 2012, 8, 1–13.
[CrossRef]
Sensors 2020, 20, 548 20 of 21

13. Jia, S.; Fadlullah, Z.M.; Kato, N.; Zhang, L. Eco-Udc: An energy efficient data collection method for disaster
area networks. In Proceedings of 2016 IEEE International Conference on Network Infrastructure and Digital
Content (IC-NIDC), Beijing, China, 23–25 September 2016; pp. 130–134.
14. Li, M.; Nishiyama, H.; Kato, N.; Owada, Y.; Hamaguchi, K. On the energy-efficient of hroughput-based
scheme using renewable energy for wireless mesh net- works in disaster area. IEEE Trans. Emerg. Top.
Comput. 2015, 3, 420–431. [CrossRef]
15. Suzuki, H.; Kaneko, Y.; Mase, K.; Yamazaki, S.; Makino, H. An ad hoc network in the sky, skymesh, for
large-scale disaster recovery. In Proceedings of IEEE Vehicular Technology Conference, Melbourne, Australia,
7–10 May 2006; pp. 1–5.
16. Sakano, T.; Kotabe, S.; Komukai, T.; Kumagai, T.; Shimizu, Y.; Takahara, A.; Ngo, T.; Fadlullah, Z.M.;
Nishiyama, H.; Kato, N. Bringing movable and deployable networks to disaster areas: Development and
field test of mdru. IEEE Netw. 2016, 30, 86–91. [CrossRef]
17. Ngo, T.; Nishiyama, H.; Kato, N.; Sakano, T.; Takahara, A. A spectrum-and energy–efficient scheme for
improving the utilization of MDRU-based disaster resilient networks. IEEE Trans. Veh. Technol. 2014, 63,
2027–2037. [CrossRef]
18. Chen, A.Y.; Peña-Mora, F.; Plans, A.P.; Mehta, S.J.; Aziz, Z. Supporting Urban Search and Rescue with digital
assessments of structures and requests of response resources. Adv. Eng. Inf. 2012, 26, 833–845. [CrossRef]
19. Hu, D.; Li, S.; Chen, J.; Kamat, V.R. Detecting, locating, and characterizing voids in disaster rubble for search
and rescue. Adv. Eng. Inf. 2019, 42, 100974. [CrossRef]
20. Peña-Mora, F.; Chen, A.Y.; Aziz, Z.; Soibelman, L.; Liu, L.Y.; El-Rayes, K.; Arboleda, C.A.; Lantz, T.S., Jr.;
Plans, A.P.; Lakhera, S.; et al. Mobile Ad Hoc Network-Enabled Collaboration Framework Supporting Civil
Engineering Emergency Response Operations. J. Comput. Civil Eng. 2010, 3, 302–312. [CrossRef]
21. Chen, A.Y.; Peña-Mora, F.; Ouyang, Y. A collaborative GIS framework to support equipment distribution for
civil engineering disaster response operations. Autom. Constr. 2011, 20, 637–648. [CrossRef]
22. Crawford, P. S.; Al-Zarrad, M.A.; Graettinger, A.J.; Hainen, A.M.; Back, E.; Powell, L. Rapid Disaster Data
Dissemination and Vulnerability Assessment through Synthesis of a Web-Based Extreme Event Viewer and
Deep Learning. Adv. Civ. Eng. 2018, 2018; doi:10.1155/2018/7258156. [CrossRef]
23. Menon, V.G.; Pathrose, J.P.; Priya, J. Ensuring Reliable Communication in Disaster Recovery Operations with
Reliable Routing Technique. Mob. Inf. Syst. 2016, 2016; doi:10.1155/2016/9141329. [CrossRef]
24. Vo, N.-S.; Masaracchia, A.; Nguyen, L.D.; Huynh, B.-C. Natural Disaster and Environmental Monitoring
System for Smart Cities: Design and Installation Insights. EAI Endorsed Trans. Indust. Netw. Intellig. Syst.
2018, 5, e5. [CrossRef]
25. Nguyen, L.D.; Kortun, A.; Duong, T.Q. An Introduction of Real-Time Embedded Optimisation Programming
for UAV Systems under Disaster Communication. EAI Endorsed Trans. Indust. Netw. Intellig. Syst. 2018, 5, e5.
[CrossRef]
26. Souryal, M.R.; Geissbuehler, J.; Miller, L.E.; Moayeri, N. Real-Time Deployment of Multihop Relays for
Range Extension. In Proceedings of the 5th international conference on Mobile systems, applications and
services. (MobiSys), San Juan, PR, USA, 11–14 June 2007; pp. 85–98. [CrossRef]
27. Wolff, A.; Subik, S.; Wietfeld, C. Performance analysis of highly available ad hoc surveillance networks
based on dropped units. In Proceedings of the 2008 IEEE Conference on Technologies for Homeland Security,
Waltham, MA, USA, 12–13 May 2008; pp. 123–128.
28. Bao, J.Q.; Lee, W.C. Rapid deployment of wireless ad hoc backbone net works for public safety incident
management. In Proceedings of the IEEE GLOBECOM 2007—IEEE Global Telecommunications Conference,
Washington, DC, USA, 26–30 November 2007; pp. 1217–1221.
Sensors 2020, 20, 548 21 of 21

29. Liu, H.; Xie, Z.; Li, J.; Li, S.; Siu, D.J.; Hui, P.; Whitehouse, K.; Stankovic, J.A. An automatic, robust, and
efficient multi-user breadcrumb system for emergency response applications. IEEE Trans. Mob. Comput.
2013, 13, 723–736. [CrossRef]
30. Khan, A.; Ullah, F.; Kaleem, Z.; Rehman, S.U.; Anwar, H.; Cho, Y.-Z. EVP-STC: Emergency Vehicle Priority
and Self-Organising Traffic Control at Intersections Using Internet-of-Things Platform. IEEE Access 2018, 6,
68242–68254. [CrossRef]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).