Ad Hoc Communication Topology Switching during Disasters

from Altruistic to Individualistic and Back

Indushree Banerjee, Martijn Warnier and Frances Brazier

Faculty of Technology Policy and Management, Systems Engineering and Simulations,

Delft University of Technology, Delft, The Netherlands

Keywords:

Ad Hoc Communication Networks, Disaster Management, Topology Switching, Autonomous Computing.

Abstract:

Disaster communication has made immense progress in the last thirty years. At present, disaster research fo-

cuses on bottom-up approaches such as civilian inclusion in disaster response. With the advent of smartphones,

citizen-based emergency communication has become possible. Present ad hoc communication technologies

typically form a fully connected mesh network, which connects all phones that are within each other’s trans-

mission range. This facilitates low-latency direct communication between citizens, but it quickly drains the

battery of phones. Alternative ad hoc communication networks form an adaptive energy-efﬁcient network

topology, that is most draining to batteries of phones that have a higher charge, while low-energy phones

are spared from relaying messages, thereby preserving battery and thus maintaining their connection with the

rescue communication network. Both of these approaches have their own advantages. Which one is best

for communication needs depends on the context. This position paper discusses the possibility of a decision

model as an approach to automatically switch between the two alternative ad hoc communication networks.

This ensures that citizens in disasters can make use of the optimal communication system at all times.

1 INTRODUCTION

To enhance societal resilience against disasters, col-

lective participation of citizens in a timely and in-

formed manner must be incorporated (Comfort and

Haase, 2006; Comfort et al., 2010). One way of facil-

itating citizen autonomy is to design emergency com-

munication services with existing tools such as smart-

phones that can provide continuous access to informa-

tion (Maryam et al., 2016; Kumbhar et al., 2016).

To form a mobile ad hoc communication net-

work (MANET) smartphones use their inbuilt WiFi or

Bluetooth to connect with other devices in their prox-

imity or transmission range to exchange messages in a

peer-to-peer mode, forming communication networks

on-the-ﬂy (Wang et al., 2017; Raffelsberger and Hell-

wagner, 2013). Ad hoc networks such as MANETs

(Mobile Ad Hoc Networks) have two main topologies

for their underlying connectivity: full mesh or scale-

free.

The topology of a network determines the pat-

tern of connectivity between nodes to form connec-

tions. In a full mesh topology, nodes within trans-

mission range of each other form a direct point-to-

point connection. This leads to a fully connected net-

work such that every node in the transmission range

gets connected and as they move around they make

more connections dynamically. If the sender and

the receiver are not in direct contact, there are al-

ways relaying nodes that can pass the message to the

intended destination. Applications such as Firechat

(Lin et al., 2015), ServalMesh (Lieser et al., 2017),

HelpMe (Mokryn et al., 2012) follow this topology

and are promoted as solutions to disaster emergency

communications.

The performance of the network depends on the

chosen topology, and each topology has its own ben-

eﬁt. A full mesh topology provides more connectiv-

ity and reliability due to redundant routes. However,

scaling a full mesh topology is a challenge. Addition-

ally, a full mesh topology can lead to extreme battery

drainage of participating nodes due to high connec-

tion costs.

Alternatively in a scale-free network topology,

like the recently developed SOS (Banerjee et al.,

2020), delay and latency can be very high. There-

fore, to cater to a speciﬁc requirement, the topology

is predetermined for a speciﬁc application.

For example, for a sparsely populated area, a

scale-free network topology is preferred. However,

Banerjee, I., Warnier, M. and Brazier, F.

Ad Hoc Communication Topology Switching during Disasters from Altruistic to Individualistic and Back.

DOI: 10.5220/0009434201030107

In Proceedings of the 5th International Conference on Complexity, Future Information Systems and Risk (COMPLEXIS 2020), pages 103-107

ISBN: 978-989-758-427-5

103

Figure 1: This ﬁgure represents two different topologies. On the left a full mesh topology is represented (individualistic

topology) and on the right a scale-free network topology is illustrated (altruistic).

for a densely populated area a full mesh topology is

chosen.

In addition to density and mobility, there are many

other factors that play a role in maintaining connec-

tivity, such as available resources, whether charging

facilities are available, context and social information

or relationships (Jedari et al., 2018). This can lead

to non-cooperative behaviour among relaying nodes.

Based on the behaviour of nodes, literature (Jedari

et al., 2018) classiﬁes mobile nodes in three main cat-

egories,

• selﬁsh

• cooperative, and

• malicious nodes.

Malicious nodes are out of the scope of this paper.

This paper mainly focuses on selﬁsh node behaviour

and cooperative node behaviour.

A selﬁsh node can on purpose drop messages ei-

ther to save resources, decline to connect due to pri-

vacy issues and may have social biases before joining

an ad hoc network in a community. This can limit the

coverage area and scaling of the network, while low-

ering reliability of message delivery due to network

segmentation.

Therefore, there is a need for networks to be

able to switch between topologies depending on the

spatial-temporal-resource context.

This paper addresses these criteria required for a

decision model that generates a self-organised topol-

ogy with respect to spatial-temporal-resource context.

The purpose is to switch between two main topolo-

gies, namely individualistic and altruistic.

In this paper a full mesh topology is called an in-

dividualistic topology since it forms connections and

routes data without consideration for other nodes in

the network. As represented on the left panel of ﬁg-

ure 1, each node connects with every other node in its

proximity or transmission range and communicates

without any consideration of relaying nodes.

During disaster where there is continued uncer-

tainty of available resources and need of continued

access, it is very possible that nodes demonstrate self-

ish behaviour and behave more individualistically.

At present three main categories of incentive

mechanisms are present that promote co-operative be-

haviour. Speciﬁc to disaster scenarios are reputation-

based, credit-based and tit-for-tat-based (Gupta et al.,

2014; Radenkovic et al., 2018; Asuquo et al., 2016).

However, these mechanisms mainly focus on data

routing and not topology creation.

Therefore, this paper deﬁnes a scale-free network

as an altruistic topology that forms an ad hoc net-

work while considering node limitations in terms of

resources. The right panel of Figure 1 shows the for-

mation of an altruistic network that promotes equal

participation of nodes to avoid low-energy nodes for

relaying, thus preventing selﬁsh behaviour.

During disasters, an effective response is highly

dependent upon the ability of a system to sustain

dynamic changes. Systems must be modeled with

the consideration of dynamic context, while updating

information and a continued access to information,

COMPLEXIS 2020 - 5th International Conference on Complexity, Future Information Systems and Risk

104

Figure 2: Decision model framework.

therefore, the decision model must allow switching

between these two connection topologies.

2 DECISION MODEL AND

DECISION CRITERIA

The switching can be determined by various factors

such as the density of nodes, mobility of nodes, avail-

ability of resources such as charging facilities, social

participation, inclusion of traditional rescue among

others.

Access to updated information needs to be con-

tinuous and thus a smooth automatic transition that

avoids discontinued communication service is very

necessary.

The ﬁrst transition takes place in the immedi-

ate aftermath of a disaster, i.e. when infrastructures

are unavailable, services switch to a bottom-up ap-

proach. This transition is reversed when infrastructure

is either restored or traditional rescue brings equip-

ment such as unmanned aerial vehicles or WiFi ac-

cess points to re-establish communication and install

a top-down approach.

The second transition takes place between topolo-

gies catering to bottom-up approaches, between altru-

istic and individualistic topologies. This framework is

represented in ﬁgure 2.

2.1 The Decision Tree

This section presents an example of decision trees that

determine when to switch between the two types of

topologies in the bottom-up approach. The decision

is based on three factors:

1. charge of the participating nodes,

2. node density, and

3. number of messages being exchanged.

First, charge of the participating nodes is central in

deciding which topology to use, since node partici-

pation is a top priority. If the charge of participating

nodes is low, low energy nodes start running out of

battery, and stop participating.

Second, node density determines how many nodes

are in transmission range, and therefore determines

how many connections need to be made. Getting con-

nected is an expensive process. If there are many

nodes in range, and every node connects to all nodes

in range, this drains the battery quickly, and will in

the end not be efﬁcient.

Third, if the number of messages being sent is

high, then a topology that relies on nodes relaying

messages via many intermediate nodes will drain bat-

tery quickly, and will not be efﬁcient.

In the proposed decision models, this last factor is

not included, because it is not a top priority and the

decision model should be simple to conserve energy

in computation.

In the decision model, a number of criteria are

formulated for evaluating whether a node can best

switch to a different type of topology, or best stays

with the present topology in use. These decision

models are not mirror images: Once an altruistic

topology is in place, the threshold should be high to

switch to an individualistic topology.

Ad Hoc Communication Topology Switching during Disasters from Altruistic to Individualistic and Back

105

Figure 3: Decision tree for switching between altruistic to individualistic and back.

The decisions as represented in ﬁgure 3 are followed

at the level of individual nodes: Every node decides

for itself whether it will switch to the different topol-

ogy, and although this decision is shared with the sur-

rounding nodes, it will only affect the connections of

this node.

Therefore, the topology of the whole network can

be a hybrid between the two types of topologies, with

some areas or some subsets of nodes following the in-

dividualistic topology, and other areas or subsets of

nodes following the altruistic topology. The decisions

are all dichotomized using thresholds, to keep com-

putations simple and low in energy consumption, as

they need to be performed by each individual phone

at regular intervals.

The switching is performed by adaptive self-

organisation that follows the principles of Autonomic

computing (Brazier et al., 2009). Autonomic com-

puting has been used for designing complex system

through self-organisation and self-management of in-

dividual entities participating in forming the system

without human intervention.

The process involves each entity monitoring the

environment to acquire context information, analyz-

ing the information to gather perspective, planning

based on the analyses and ﬁnally executing the de-

cision. The same decision criteria will be followed by

each node.

In the ad hoc networks, participating nodes are not

aware of the context and their own limitations at the

beginning as they join the network. As the network

formation begins and the number of nodes participat-

ing increases, the context changes and thus they need

to change their connectivity pattern to maintain com-

munication and coverage.

The approach includes setting thresholds for var-

ious decision criteria. In the extended version of this

paper, the algorithms for switching will be extended.

3 CONCLUSION

In conclusion this paper presents the conceptual

framework of a decision model that allows topol-

ogy switching using autonomic computing and self-

organisation for emergency communications in the af-

termath of a disaster.

The purpose is to allow smooth transition of dif-

ferent connectivity patterns that allows human col-

lective intelligence to be utilized via technical means

to support its society in complex, dynamic environ-

ments. An adaptive system capable of facilitating

communication between affected citizens provides

citizens the autonomy to help themselves.

REFERENCES

Asuquo, P., Cruickshank, H., Ogah, C. P. A., Lei, A.,

and Sun, Z. (2016). A collaborative trust manage-

ment scheme for emergency communication using de-

lay tolerant networks. In 2016 8th Advanced Satel-

lite Multimedia Systems Conference and the 14th Sig-

nal Processing for Space Communications Workshop

(ASMS/SPSC), pages 1–6. IEEE.

Banerjee, I., Warnier, M., and Brazier, F. M. T. (2020).

Self-organizing topology for energy-efﬁcient ad-hoc

communication networks of mobile devices. Complex

Adaptive Systems Modeling. Journal article under re-

view (preprint available online).

Brazier, F. M., Kephart, J. O., Parunak, H. V. D., and Huhns,

M. N. (2009). Agents and service-oriented computing

for autonomic computing: A research agenda. IEEE

Internet Computing, 13(3):82–87.

Comfort, L. K. and Haase, T. W. (2006). Communication,

coherence, and collective action: The impact of hurri-

COMPLEXIS 2020 - 5th International Conference on Complexity, Future Information Systems and Risk

106

cane katrina on communications infrastructure. Public

Works Management & Policy, 10(4):328–343.

Comfort, L. K., Oh, N., Ertan, G., and Scheinert, S. (2010).

Designing adaptive systems for disaster mitigation

and response. Designing resilience: Preparing for ex-

treme events, pages 33–62.

Gupta, A. K., Bhattacharya, I., Banerjee, P. S., and Mandal,

J. K. (2014). A co-operative approach to thwart selﬁsh

and black-hole attacks in dtn for post disaster scenario.

In 2014 Fourth International Conference of Emerging

Applications of Information Technology, pages 113–

118. IEEE.

Jedari, B., Xia, F., and Ning, Z. (2018). A survey on human-

centric communications in non-cooperative wireless

relay networks. IEEE Communications Surveys & Tu-

torials, 20(2):914–944.

Kumbhar, A., Koohifar, F., G

uvenc¸, I., and Mueller, B.

(2016). A survey on legacy and emerging technolo-

gies for public safety communications. IEEE Com-

munications Surveys & Tutorials, 19(1):97–124.

Lieser, P., Alvarez, F., Gardner-Stephen, P., Hollick, M.,

and Boehnstedt, D. (2017). Architecture for respon-

sive emergency communications networks. In 2017

IEEE Global Humanitarian Technology Conference

(GHTC), pages 1–9. IEEE.

Lin, W. Y., Hsueh, K.-P., and Pa, P.-S. (2015). The de-

velopment of emergency communication app using ad

hoc network with ipv6. In 2015 International Confer-

ence on Intelligent Information Hiding and Multime-

dia Signal Processing (IIH-MSP), pages 41–44. IEEE.

Maryam, H., Shah, M. A., Javaid, Q., and Kamran, M.

(2016). A survey on smartphones systems for emer-

gency management (spsem). International Journal

of Advanced Computer Science and Applications,

7(6):301–311.

Mokryn, O., Karmi, D., Elkayam, A., and Teller, T. (2012).

Help me: Opportunistic smart rescue application and

system. In 2012 The 11th Annual Mediterranean Ad

Hoc Networking Workshop (Med-Hoc-Net), pages 98–

105. IEEE.

Radenkovic, M., Walker, A., and Bai, L. (2018). To-

wards better understanding the challenges of reliable

and trust-aware critical communications in the after-

math of disaster. In 2018 14th International Wire-

less Communications & Mobile Computing Confer-

ence (IWCMC), pages 648–653. IEEE.

Raffelsberger, C. and Hellwagner, H. (2013). A hybrid

manet-dtn routing scheme for emergency response

scenarios. In 2013 IEEE international conference on

pervasive computing and communications workshops

(PERCOM workshops), pages 505–510. IEEE.

Wang, Y., Wei, L., Vasilakos, A. V., and Jin, Q. (2017).

Device-to-device based mobile social networking in

proximity (msnp) on smartphones: Framework, chal-

lenges and prototype. Future Generation Computer

Systems, 74:241–253.

Ad Hoc Communication Topology Switching during Disasters from Altruistic to Individualistic and Back

107