Author: Rohan de Silva PhD, SMIEEE
CQUniversity Sydney Campus
Apart from being flown as hobby devices, micro and mini Unmanned Aerial Vehicles (UAVs)
or drones have found their way into several commercial application domains such as
agriculture, livestock, natural disaster, parcel delivery and surveillance. However, it is only
individual UAVs controlled by either RC controllers or smartphones and tablets but not UAV
networks that are used in these applications.
Instead of using individual UAVs, if suitable UAV networks could be designed, some of these
commercial applications can be executed more efficiently and reliably. A UAV network is, in
essence, a communication network of UAVs where each UAV acts as a communication
node. Just as in the case of a single UAV, a UAV network also needs at least one ground
station (GS) such as a smartphone or a computer that sends control messages to and
receives navigation information as well as video streams from the UAVs.
There are several topologies for UAV networks that have been proposed by various authors.
In direct topology, one GS communicates directly with the UAVs but the UAVs cannot
communicate with each other without the help of the GS. In the satellite topology, instead
of the GS directly communicating with each UAV, it uses a satellite in between. The cellular
topology makes use of a cellular infrastructure network comprising of a number of base
stations such as in a cellular mobile network. A UAV that is in a particular cell of the cellular
network communicates with the corresponding base station. In the mesh topology, each
UAV can communicate with a base station as well as with each other. This network
topology is similar to the topology of a Mobile Ad Hoc Network (MANET).
A Flying Ad Hoc Network (FANET) is essentially a MANET where all the mobile devices are
UAVs. Being an ad hoc network, the communication paths in a FANET are determined using
ad hoc routing protocols that are designed for MANETs and Vehicular Ad Hoc Networks
In order any ad hoc routing protocol to work properly there should be neighbour nodes. In
a FANET since the UAVs can fly arbitrarily to undertake their tasks, it is possible that a UAV
node may not find another UAV neighbour node to route the packets. However, as there
are a large number of UAV nodes belonging to different individuals or organisations, this is
unlikely to happen. Nevertheless, it is quite different in some of the commercial
applications where the UAVs do belong to one individual or organisation. In such an
environment, it is not possible to have a large number of UAVs due to their cost and
maintenance. If an organisation decides to use its UAVs in a FANET together with UAVs
belonging to others to limit the required number of its UAVs, then the security of its
information will be threatened.
An Internet of Drones (IoD) has a cellular topology where the geographical area is divided
into cells, each consisting of a base station. These base stations are connected together
with wired or wireless links, and when a UAV flies over a cell, its connection is handed off to
the base station of that cell. The advantage of an IoD is that the UAVs can be controlled via
the existing 4G or 5G cellular mobile network, though this is costly and the control
performance would be poor due to delay. Besides, 4G or 5G is not available in remote areas
of some countries where some of the application scenarios take place. A cellular
infrastructure is also an issue in natural disaster situations where most of the
infrastructures would be destroyed.
After considering these issues of FANETs and IoDs, the author conceived the idea of
creating a special UAV network for these commercial applications in 2013. The concept was
to have a mesh network of UAVs with a single GS but, instead of using ad hoc routing as in
FANETs, to facilitate the communication using switching in the Media Access Control (MAC)
layer, and control the movements of UAVs such that there is always a communication path
from each UAV to the ground station. This network was called a private UAV network as it
was designed with the aim of commercial applications involving UAVs of one individual or
an organization. The UAV network research at CQUniversity Sydney commenced in 2015
with the joining of the postgraduate research student, Prabhu Jyot Singh and it was later
identified that several other practical issues have to be resolved in order to develop a
private UAV network that is suitable to be employed in these application scenarios.
One of them was the number of UAVs required in the mesh topology. The author noticed
that, in private UAV networks, there could be several UAVs that do not take part in
communication at various situations and hence, statistically, there is a waste of resources.
In October 2017, a second research student, Sandaruvan Rajasinghege joined the UAV
network research group of the author and many discussions took place as to how to resolve
this issue. In late 2017, the author put forward the idea of LANs of Drones (LoDs) and
proposed a special topology called star-connected relay topology. As shown in Fig. 1, in this
topology, groups of UAV nodes are connected at tandem called branches. The UAV node at
the end of each branch that is closest to the GS and in its Wi-Fi range is connected to the
Fig. 1: A LAN of Drones -
Please download the full article