RescueMe: Smartphone-Based Self Rescue System

for Disaster Rescue

Cong Pu Xitong Zhou
Weisberg Division of Computer Science MS Graduate of Computer Science
Marshall University Marshall University
Huntington, WV 25755 Huntington, WV 25755
puc@marshall.edu zhou34@marshall.edu

Abstract—Recent ubiquitous earthquakes have been leading communication infrastructures were physically damaged or
to mass destruction of electrical power and cellular infrastruc- lacked the energy necessary to operate. [3]. More importantly,
tures, and deprive the innocent lives across the world. Due since many people could be trapped beneath the rubble and
to the wide-area earthquake disaster, unavailable power and
communication infrastructure, limited man-power and resource, brick, the victims may have a large chance to survive if
traditional rescue operations and equipment are inefficient and they are located and rescued within ”Golden 72 Hours”. If
time-consuming, leading to the golden hours missed. With a severely injured person does not receive care or medical
the increasing proliferation of powerful wireless devices, like treatment quickly, the probability of survival rapidly decreases.
smartphones, they can be assumed to be abundantly available Thus, in order to minimize casualties and save innocent lives
among the disaster victims and can act as valuable resource to
coordinate disaster rescue operations. In this paper, we propose across the world, it is essential to plan and conduct expedited
a smartphone-based self rescue system, also referred to as disaster rescue operations.
RescueMe, to assist the operations of disaster rescue and relief. However, when an earthquake occurs, the rescue teams or
The basic idea of RescueMe is that a set of smartphones carried planners of disaster rescue and relief mainly suffer from the
by survivors trapped or buried under the collapsed infrastructure following issues. First, since the current disaster situation of
forms into a one-hop network and sends out distress signal
in an energy-efficient manner to nearby rescue crews to assist sudden earthquake may not be available, it is difficult to make
rescue operations. We evaluate the proposed approach through a plan or decision on the priority of rescue operations in
extensive simulation experiment and compare its performance terms of the focused rescue areas, the distribution of rescue
with the existing scheme TeamPhone. The simulation results show teams, or the allocation of equipment. Second, the impact
that the proposed approach can significantly reduce the schedule areas are admittedly wide, ranging from a few miles to several
vacancy of broadcasting distress signal and improve the discovery
probability with very little sacrifice of network lifetime, and U.S. states, but the number of rescue teams and man-power
indicate a potentially viable approach to expedite disaster rescue are very limited in reality. Third, due to the collapse of
and relief operations. power and communication infrastructure, the impacted areas
become a blackspot where WiFi and 4G-LTE services are not
Index Terms—Natural Disaster, Disaster Rescue and Relief, available and the affected area is cut off from the outside.
Scheduling, Distress Signal, Smartphone
Last but not least, most rescue teams still heavily rely on the
traditional operations and equipment, such as detection dogs,
I. I NTRODUCTION
video cameras, or sound sensors. In summary, the traditional
Unexpected natural disasters such as tornadoes, earthquakes, rescue operations and equipment are inefficient and time-
hurricanes, and tsunamis have been rising dramatically in consuming, leading to the golden hour missed.
recent years. In particular, earthquakes tremendously kill in- On the other hand, smartphones have become an essential
nocent lives and damage the environment around the globe, electronic device what people always carry for communication
and the epicenter of an earthquake can occur anywhere and and social connection, or place them where they can be easily
now no place would be safe from ubiquitous earthquakes. For and immediately accessed. It has been predicted that the
example, a 5.6-magnitude earthquake struck Oklahoma and number of smartphone users in the United States is estimated
impacted six neighboring states in the U.S. on September 04, to reach 148.68 million in 2019, and the number of smartphone
2016 [1]. The Ecuador earthquake (April 16, 2016) left a 272 users worldwide forecasts to exceed 2 billion by that time [4].
death toll and more than 2,500 injured [2]. An earthquake often Thus, with the increasing proliferation of smartphones, they
happens in a flash but has the potential to massively destruct can be assumed to be abundantly available among the disaster
the infrastructures, buildings, and homes in a short period of victims and act as valuable resources to coordinate rescue
time. After the disaster, it was impossible for disaster victims operations. For example, after the Haiti earthquake in 2010,
to utilize their communication devices, such as smartphone, there were approximately 2.8 million active mobile subscribers
tablet, or laptop, to notify their families and friends of their out of 10 million inhabitants contributing data for tracking the
safety and confirm the safety of their loved ones since the movement of population in the affected region [5].

978-1-7281-0554-3/19/$31.002019
c IEEE
 In light of these, this paper is to develop a novel disaster share information among people gathered in evacuation cen-
self rescue system incorporating with increasingly popular ters. In the [10], a robot snakes with hyper-redundant body and
smartphone in the disaster area. Our major contribution is unique gaits is proposed to offer a promising solution to search
summarized in the following: and rescue applications in the disaster. An unmanned aerial
• We propose a smartphone-based self rescue system, also vehicles-aided disaster rescue system is proposed to locate
referred to as RescueMe, to assist the operations of possible victims in [11], where unmanned aerial vehicles [12]
disaster rescue and relief. The basic idea of RescueMe fly around the disaster area and sniff out wireless signals from
is that a set of smartphones carried by survivors trapped any mobile devices to support the search team to narrow down
or buried under the collapsed infrastructure forms into the search area within meters.
a one-hop network and sends out distress signal in an The [13] presents a smartphone-assisted victim localization
energy-efficient manner to nearby rescue crews to assist method in which smartphones belonging to trapped victims
rescue operations. and other people in disaster affected areas can self-detect
• We develop a customized simulator framework and im- the occurrence of a disaster incident by monitoring the radio
plement the proposed scheme for experiment study. For environment and can self-switch to a disaster mode to transmit
performance comparison, we revisit an existing approach emergency help messages with their location coordinates to
TeamPhone [6], and modify it to work in the developed other smartphones nearby. To locate other neighboring smart-
simulation framework. phones also operating in the disaster mode, each smartphone
We compare and analyze the performance of RescueMe and runs a rendezvous process. In the [14], an application, also
TeamPhone in terms of network lifetime and schedule vacancy referred to as SOSCast, is proposed to propagate SOS mes-
of broadcasting distress signal through extensive simulation sages from trapped survivors through a direct communication
experiments. The simulation results show that the proposed between smartphones. By collecting SOS messages that in-
approach can significantly reduce the schedule vacancy of clude significant information such as their name, state, and
broadcasting distress signal and improve the discovery proba- location, rescuers can estimate the locations of the survivors.
bility with very little sacrifice of network lifetime, and indicate Without relying on any infrastructure, the [15] presents a
a potentially viable approach to expedite disaster rescue and new algorithm that allows the smartphones of the rescuers
relief operations. and victims to seamlessly collaborate in order to estimate the
The rest of paper is organized as follows. The prior ap- locations of the victims by using both the received signal
proaches are summarized and analyzed in Section II. The strength indicator of the WiFi signals and the GPS information
proposed smartphone-based self rescue system is presented in of the rescuers’ smartphones. The [16] proposes a smartphone
Section III. Section IV presents extensive simulation results and IoT devices-assisted emergency and recovery method in a
and analyses. Finally, we conclude the paper in Section V. post-disaster environment, where smartphones can utilize the
IoT devices in the disaster affected areas to successfully relay
II. R ELATED W ORK the emergency messages to other smartphones.
Using smartphones for constructing disaster recovery net- In [6], by bridging the gap among different kinds of
work and assisting the operations of disaster rescue and relief wireless networks, a system called TeamPhone is proposed
has been extensively explored in the last decade. to provide smartphones the capabilities of communications
The [7] develops a mechanism to enable the devices to in disaster recovery. TeamPhone consists of two components:
discover their neighbors autonomously and transmit data of a messaging system and a self-rescue system. The messag-
disaster-affected area by different network to WiFi access ing system integrates cellular networking, ad-hoc networking
points using a smartphone-based WiFi tethering technique. In and opportunistic networking seamlessly, and enables com-
the [8], a novel architecture called energy aware disaster re- munications among rescue workers. The self-rescue system
covery network using WiFi tethering is proposed to create the energy-efficiently groups the smartphones of trapped survivor
desired network infrastructure using wireless device. The basic and sends out emergency messages so as to assist rescue
idea is to make use of WiFi tethering technology ubiquitously operations. However, the self-rescue system does not consider
available on wireless devices, like smartphones and tablets, that each smartphone of trapped survivor may carry different
to set up an ad hoc network for data collection in disaster amount of residual energy, and the smartphone with less
scenarios. The [9] proposes a smartphone-based post-disaster residual energy may turn off quickly because of frequently
management mechanism in the disaster affected areas using the broadcasting emergency messages. Thus, the schedule of send-
concepts of WiFi tethering, where smartphones in the affected ing out emergency messages should be dynamically adjusted
areas may turn themselves into temporary WiFi hotspots to accordingly when the network topology changes because of
provide Internet connectivity and important communication the death of certain smartphone.
abilities to nearby WiFi-enabled user devices. In the [3], the In summary, various kinds of smartphone-based wireless
concept of multihop device-to-device communication network communication technologies and hybrid networks have been
systems integrating with different wireless technologies is widely investigated for disaster rescue and relief. However, to
proposed to deliver messages using only users mobile devices, the best of authors’ knowledge, the proposed research focusing
send out emergency messages from disconnected areas, and on dynamically adjusting the schedule of sending out distress
     

       
    



 
     
 
 


   



   
     


   




    
  
     
(a) Broadcast Hello Messages (b) One-hop Neighbor List (c) One-hop Network

Fig. 1. A set of nodes broadcast Hello messages to build one-hop neighbor list G∗ and one-hop network.

signal according to the change of network topology is new. constructs an one-hop network. The one-hop network consists
of cliques, in which the direct connection exists between every
III. T HE P ROPOSED S MARTPHONE -BASED S ELF R ESCUE two nodes, and each node can communicate with every other
S YSTEM node. In order to extend the battery lifetime through reducing
In this section, we first present an overview of the proposed energy consumption, the nodes in the one-hop network wake
smartphone-based self rescue system, also referred to as Res- up alternatively in a coordinated way and send out distress
cueMe, and then discuss the RescueMe and its corresponding signal to discover nearby rescue crew. In addition, each node
techniques in details. may be equipped with different amount of battery energy
when the disaster happens, and the node with less amount
A. Overview of the RescueMe
of energy will turn off quickly. Thus, the wake-up schedule of
With the proliferation of powerful smartphones, they can broadcasting distress signal should be dynamically adjusted
be assumed to be abundantly available among the trapped in response to the change of one-hop network topology. In
survivors in the affected region and act as valuable resources. the following, we investigate three major issues to implement
Most smartphones are equipped with both WiFi and Bluetooth the RescueMe: (i) how to construct the one-hop network;
transceivers. WiFi technology allows a smartphone to connect (ii) how to determine the wake-up schedule of broadcasting
to a wireless local area network (WLAN) based on the IEEE distress signal; and (iii) how to dynamically adjust the wake-
802.11 standard. WiFi running in the 2.4 GHz radio band up schedule according to the change of network topology.
covers a larger range than 5.0 GHz. Unlike WiFi consuming
high battery power, Bluetooth wireless technology (BWT) B. RescueMe: Smartphone-Based Self Rescue System
is primarily designed for a short range communication with First, each node broadcasts an one-time Hello message
low-power consumption. With the support of WiFi or BWT, piggybacked with its node id, overhears Hello messages broad-
survivors even trapped under the rubble can periodically send casted by other nodes, and then build the one-hop neighbor
out a distress signal using their smartphone. If each smart- list G∗ [17], [18]. For example, as shown in Subfig. 1(a), nm
phone continuously stays awake and sends out distress signal, broadcasts a Hello message, and overhears the Hello messages
this will increase discoverability and reachability of trapped broadcasted by its adjacent nodes (e.g., nh , ni , ng , nk , nl , and
survivors. However, continuously broadcasting distress signal nn ). As a result, nm can build its one-hop neighbor list, G∗m
can quickly drain battery, and rescue operations may even last = {h,i,g,k,l,n}, as shown in Subfig. 1(b). By using the same
for days after disasters occur. Thus, how to collaborate smart- technique, other nodes also can build their one-hop neighbor
phones of survivors trapped under the collapsed infrastructure list. In this paper, each node also considers itself as an one-
to send out distress signal in an energy-efficient manner is a hop neighbor node and adds its id in the one-hop neighbor
challenge problem. list. Thus, G∗m = {h,i,g,k,l,n,m}. Each node then exchanges its
In this paper, we propose a smartphone-based self rescue one-hop neighbor list G∗ with adjacent nodes, and identifies
system, also referred to as RescueMe, to assist the operations the center node who has the largest number of neighbor nodes
of disaster rescue and relief in the disaster area. The basic and the G∗ of all other nodes is a subset of the center node’s
idea of RescueMe is that a set of smartphones (later nodes) one-hop neighbor list. Then the center node builds a one-hop
carried by survivors trapped or buried under the collapsed network, where every other node can directly communicate
infrastructure forms into a one-hop network and sends out with the center node, vice versa. For example, as shown in
distress signal in an energy-efficient manner to nearby rescue Subfig. 1(c), an one-hop network is built by the center node
crews to assist rescue operations. After detecting seismic ng .
signal, the nodes can automatically enter into the self-rescue Second, the center node examines the received G∗ of all
mode. Or after disaster occurs, the trapped survivors can click other nodes in the one-hop network, and groups the nodes
a rescue app to enable the nodes to enter into the self-rescue that have the same subset of G∗ into a clique, where a direct
mode. In the self-rescue mode, each node broadcasts an one- connection exits between every two nodes and each node can
time Hello message, overhears bypassing Hello messages, and cover all other nodes in a clique. For example, in Subfig.
 {

 
  {


  {

 


   
   
   

 {

 

     {

 

   


  
  
  

 


                
{


   
  {


(a) A Snapshot of Three Cliques (b) Schedule of Sending Out Distress Signal (c) Combine Cliques

Fig. 2. A snapshot of three cliques in one-hop network, schedule of sending out distress signal, and combine existing cliques.

9
100 the cliques is 20 and 14 . Thus, nm will broadcast distress signal
earlier than nh . The center node determines the broadcasting
80 schedule for all the nodes based on the following two criteria:
Change of Coverage (%)

(i) the broadcasting frequency for each node must be ψ; (ii)

60 there is no more than one node broadcasting distress signal
at any time in the clique. After determining the schedule,
40 the center node broadcasts the Schedule packet, pktsched ,
piggybacked with the broadcasting schedule to all other nodes
20 in the one-hop network. When the node successfully receives
the pktsched packet, it replies an Ack packet, pktAck , back to
0 the center node after a randomly backoff period [19]. Here,
1 2 3 4 5 6 7
Number of Node In The Clique Subfig. 2(b) shows the schedule of sending out distress signal
of each node in the one-hop network of Subfig. 2(a), where
Fig. 3. The change of coverage ratio against the number of nodes in the
clique. the schedule begins at time 0.
Fourth, since each node may carry different amount of
1(b), ng , ni , nh , and nm contain the same subset of one- residual energy, and the node with less residual energy may
hop neighbor list, G∗sub = {i,h,m,g}. Thus, the center node turn off quickly because of frequently broadcasting distress
ng groups ni , nh , and nm along with itself into a clique, signal, thus, the schedule of broadcasting distress signal
C {i,h,m,g} , as shown in Subfig. 2(a). In the one-hop network should be dynamically adjusted in response to the change
shown in Subfig. 2(a), there are another two cliques, which of network topology. If the node is about to run out of the
are C {m,n,l,k,g} and C {j,k,g} , respectively. In a clique, nodes battery energy, it will notify the center node by sending a
are close to each other, thus, the area covered by one node TurnOff packet, pktof f . As a result, the center node removes
can be a larger proportion of the area covered by all nodes in the leaving node from existing cliques, and rebroadcasts the
the clique. Here, Fig. 3 shows the change of coverage ratio pktof f packet to all other nodes in the one-hop network.
between the coverage of one node and the total coverage of After receiving the pktof f packet, each node also removes
all nodes in a clique against the number of nodes in a clique. the leaving node from its one-hop neighbor node list. Then,
Third, since the center node belongs to multiple cliques the center node examines the updated cliques, combines the
in the one-hop network, it will determine the schedule of cliques if one is the subset of another, determines a new
broadcasting distress signal for all nodes across all the cliques. broadcasting schedule, and broadcasts the new schedule to all
In order to reduce the energy consumption, the node that is other nodes. For example, as shown in Subfig. 2(c), node nj
located in the multiple cliques will choose the lowest fraction leaves the network and the center node ng combines C {j,k,g}
of the clique that it belongs to as its broadcasting frequency, and C {m,n,l,k,g} into C {m,n,l,k,g} . If the center node has to
denoted as ψ. For example in Subfig. 2(a), the center node leave the network, it will broadcast a Dismiss packet, pktdis ,
ng belongs to three cliques, C {i,h,m,g} , C {m,n,l,k,g} , and to all other nodes. After receiving the pktdis packet, all other
C {j,k,g} , and the clique fraction of ng is 14 (since there are four nodes will restart RescueMe scheme from the beginning, and
nodes in C {i,h,m,g} ), 15 and 13 in the C {i,h,m,g} , C {m,n,l,k,g} , the same aforementioned operations will be applied.
and C {j,k,g} , respectively. Thus, the center node ng choose 15
as its broadcasting frequency, ψg = 15 . In addition, the center IV. P ERFORMANCE E VALUATION
node also calculates the sum of fractions of the cliques for each We develop a customized simulation framework using Java
node, and determines the schedule of broadcasting distress to conduct our experiments. The number of nodes in the
signal accordingly. In this paper, the node with the larger network is from 1 to 9, and the time interval of broadcasting
sum of fractions of the cliques will broadcast distress signal distress signal is 5 seconds. In this paper, we measure the
earlier. For example, node nm and nh has the fractions of the performance in terms of network lifetime and schedule va-
cliques { 14 , 15 } and { 14 }, respectively, the sum of fractions of cancy by changing key simulation parameters, including the
 600 1800
TeamPhone, =80% TeamPhone, =80%
TeamPhone, =60% 1600 TeamPhone, =60%
500 RescueMe, =80% RescueMe, =80%
RescueMe, =60% 1400 RescueMe, =60%
Network Lifetime

Network Lifetime
400
1200

300 1000

800
200
600
100
400

0 200
1 3 5 7 9 1 2 3 4 5
Number of Nodes Number of Broadcasted Signals in Every Interval
(a) (a)

800 1500
TeamPhone, =80% TeamPhone, =80%
700 TeamPhone, =60% 1350 TeamPhone, =60%
RescueMe, =80% RescueMe, =80%
1200
600 RescueMe, =60% RescueMe, =60%
1050
Vacancy Time

Vacancy Time
500
900
400 750

300 600

450
200
300
100
150
0 0
1 3 5 7 9 1 2 3 4 5
Number of Nodes Number of Broadcasted Signals in Every Interval
(b) (b)
Fig. 4. The performance of network lifetime and schedule vacancy against Fig. 5. The performance of network lifetime and schedule vacancy against
the number of nodes. the number of broadcasted distress signals in each interval.

number of nodes, the number of broadcasted distress signal in is out of battery, the structure of network will be changed
each time interval, and the number of cliques. For performance and other nodes will fill in the empty interval of broadcasting
comparison, we compare the proposed scheme RescueMe with distress signal, resulting in a lower schedule vacancy time.
TeamPhone. Second, we measure the network lifetime and schedule
First, we measure the network lifetime and schedule va- vacancy by varying the number of broadcasted distress signals
cancy time by varying the number of nodes in Fig. 4. In in each interval in Fig. 5. Overall, the network lifetime de-
Subfig. 4(a), the overall network lifetime of RescueMe and creases as the number of broadcasted distress signals increases
TeamPhone increase as the number of nodes increases. This in Subfig. 5(a). With the number of broadcasted distress
is because more nodes exist in the network, and the time signals in each interval increases, the energy consumption
interval of broadcasting distress signal for each node increases, in each interval increases. However, the RescueMe shows
resulting in a longer network lifetime. As the initial battery a slightly lower network lifetime than that of TeamPhone
power of each node is increasing, the longer network lifetime because the TeamPhone does not reschedule the interval of
will be observed. This is because the more power the node has, broadcasting distress signal whenever the node leaves the
the longer time the node lasts, which results in a higher net- network. As shown in Subfig. 5(b), the overall vacancy time
work lifetime. However, the RescueMe shows a slightly lower of the RescueMe and TeamPhone decreases as the number of
network lifetime than that of TeamPhone because TeamPhone broadcasted distress signals increases in every interval. The
uses the same schedule of broadcasting distress signal even RescueMe shows a lower increment of schedule vacancy time
if the node leaves the network. In Subfig. 4(b), the overall than that of the TeamPhone. This is because if the node is
schedule vacancy time of RescueMe and TeamPhone increase out of power, the RescueMe reschedules the rest of nodes in
as the number of nodes increases. This is because the schedule the network, which results in a lower vacancy time. Although
of broadcasting distress signal of TeamPhone does not change the number of broadcasted distress signal in every interval
when the nodes leave the network because of running out decreases and the network lifetime increases, the RescueMe
of battery. As a result, the schedule vacancy time increases still shows much smaller schedule vacancy time than that of
significantly. The RescueMe shows a lower schedule vacancy TeamPhone.
time than that of TeamPhone. In the RescueMe, when the node Third, Fig. 6 shows the network lifetime and schedule va-
 crews to assist rescue operations. We evaluated the proposed
650
600
TeamPhone, =80% approach through extensive simulation experiments and com-
TeamPhone, =60%
550 RescueMe, =80%
pared its performance with the existing scheme TeamPhone.
500 RescueMe, =60% The simulation results showed that the proposed approach
450 can significantly reduce the schedule vacancy of broadcasting
Network Lifetime

400
350
distress signal and improve the discovery probability with very
300 little sacrifice of network lifetime, and indicate a potentially
250 viable approach to expedite disaster rescue operations.
200
150 ACKNOWLEDGMENT
100
This research was supported by Startup grant in the Weis-
50
0
berg Division of Computer Science and 2018 John Marshall
2 3 4 5 6 University Summer Scholars Awards at Marshall University.
Number of Cliques
(a)
R EFERENCES
500 [1] C. Pu, “Evacuation Assisting Strategies in Vehicular Ad Hoc Networks,”
TeamPhone, =80%
TeamPhone, =60%
in IEEE Proc. UEMCON, November 2018.
RescueMe, =80% [2] Ecuador Earthquake: Death Toll Jumps to 272; More Than 2,500
400 Injured, https://www.cnn.com/2016/04/17/americas/ecuador-deadly-
RescueMe, =60%
earthquake, Apr 18, 2016.
Vacancy Time

[3] H. Nishiyama, M. Ito, and N. Kato, “Relay-by-Smartphone: Realizing

300
Multihop Device-to-Device Communications,” IEEE Commun. Mag.,
vol. 52, no. 4, pp. 56–65, 2014.
200
[4] Number of smartphone users in the United States from 2010
to 2022, 2018, https://www.statista.com/statistics/201182/forecast-of-
smartphone-users-in-the-us.
100 [5] L. Bengtsson, X. Lu, A. Thorson, R. Garfield, and J. V. Schreeb,
“Improved Response to Disasters and Outbreaks by Tracking Population
Movements with Mobile Phone Network Data: A Post-Earthquake
0 Geospatial Study in Haiti,” PLoS medicine, vol. 8, no. 8, 2011.
2 3 4 5 6 [6] Z. Lu, G. Cao, and T. L. Porta, “Networking Smartphones for Disaster
Number of Cliques Recovery,” in Proc. IEEE PerCom, 2016, pp. 1–9.
(b) [7] A. Pal and K. Kant, “E-Darwin2: A Smartphone Based Disaster Recov-
Fig. 6. The performance of network lifetime and schedule vacancy time ery Network using WiFi Tethering,” in Proc. IEEE CCNC, 2018, pp.
against the number of cliques. 1–5.
[8] M. Raj, K. Kant, and S. Das, “E-DARWIN: Energy Aware Disaster
cancy of the RescueMe and TeamPhone with varying number Recovery Network using WiFi Tethering,” in Proc. IEEE ICCCN, 2014,
of cliques in the network. As shown in Subfig. 6(a), the overall pp. 1–8.
network lifetime decreases as the number of cliques increases. [9] S. Ray, R. Sinha, and S. Ray, “A Smartphone-based Post-Disaster
Management Mechanism Using WiFi Tethering,” in Proc. IEEE ICIEA,
This is because when the number of cliques increases but 2015, pp. 966–971.
the total number of nodes in the network is not changed, [10] Y. Luo, J. Liu, Y. Gao, and Z. Lu, “Smartphone-Controlled Robot Snake
more nodes broadcast distress signals in every interval. As a for Urban Search and Rescue,” in Proc. ICIRA, 2014, pp. 352–363.
[11] Y. Ho, Y. Chen, and L. Chen, “Krypto: Assisting Search and Rescue
result, the energy consumption of each interval increases, and Operations Using Wi-Fi Signal with UAV,” in Proc. ACM DroNet, 2015,
the network lifetime decreases accordingly. The TeamPhone pp. 3–8.
shows a higher network lifetime than that of RescueMe, this [12] C. Pu, “Jamming-Resilient Multipath Routing Protocol for Flying Ad
Hoc Networks,” IEEE Access, vol. 6, pp. 68 472–68 486, 2018.
is because the TeamPhone does not reschedule the rest of [13] A. Hossain, S. Ray, and R. Sinha, “A Smartphone-assisted Post-Disaster
nodes for broadcasting distress signals when certain nodes are Victim Localization Method,” in Proc. HPCC/SmartCity/DSS, 2016, pp.
out of power, which results in the higher schedule vacancy 1173–1179.
[14] J. Zamora, N. Suzuki, and S. Kashihara, “SOS Message Distribution for
time. In Subfig. 6(b), as the number of cliques increases, Searching Disaster Victims,” Smartphones from an Applied Research
the overall schedule vacancy time increases. However, the Perspective, InTech, 2017.
RescueMe achieves a better performance than TeamPhone. [15] A. Al-Sadi and F. Awad, “Smartphone-Assisted Location Identification
Algorithm for Search and Rescue Services,” in Proc. IEEE ICICS, 2017,
This is because the RescueMe reschedules the network to fill pp. 276–281.
in the empty time interval, which results in the lower vacancy [16] M. Hossain and S. Ray, “A Smartphone-Assisted Device-to-Device
time. Communication for Post-disaster Recovery,” in Proc. SmartGIFT, 2018,
pp. 11–20.
[17] C. Pu, S. Lim, C. Jinseok, and J. Byungkwan, “Active Detection
V. C ONCLUSION in Mitigating Routing Misbehavior for MANETs,” Wireless Network
In this paper, we proposed a smartphone-based self rescue (2017), https://doi.org/10.1007/s11276-017-1621-z.
[18] C. Pu, S. Lim, B. Jung, and J. Chae, “EYES: Mitigating Forwarding
system to assist the operations of disaster rescue and relief. Misbehavior in Energy Harvesting Motivated Networks,” Elsevier Com-
The basic idea of RescueMe is that a set of smartphones puter Communications, vol. 124, pp. 17–30, 2018.
carried by survivors trapped or buried under the collapsed [19] C. Pu, T. Gade, S. Lim, M. Min, and W. Wang, “Lightweight Forwarding
Protocols in Energy Harvesting Wireless Sensor Networks,” in IEEE
infrastructure forms into a one-hop network and sends out Proc. MILCOM, 2014, pp. 1053–1059.
distress signal in an energy-efficient manner to nearby rescue