Page 78 - ITU Journal Future and evolving technologies – Volume 2 (2021), Issue 2

ITU Journal on Future and Evolving Technologies, Volume 2 (2021), Issue 2

Finally, autonomous agents, like drones, unnamed 3.2.2 Technical speci icity
agent boats or robots, are able to go where humans
cannot and cover rapidly an area to ind missing people There are numerous sensors on the ield that can

or assess the situation [36, 34]. Figure 2 represents an be on the ground sensors or wearable [42]. Sensors are

example of a disaster management architecture. worn by soldiers for health monitoring. They can also

be on weapons to monitor their status [40]. Like disas‑

Network speci ications In edge computing, there

ter management application, drones [41] or robots [44]

are two main channels of communication: on one part

may be used. In battle ield health monitoring, the wire‑
the communication between end devices and the edge,

less devices worn by soldiers form a Body Area Network

the other part is the connection between the edge and

(BAN) [40]. Devices communicate with each other and

the cloud. The regular network can be damaged

with the edge with the LoWPAN wireless network. These
rendering the connection between the edge and the
cloud disrupted or unstable [16, 34]. For the devices send raw data to edge networking devices which

connection between the edge and end devices, 4G and transmit it to the semantic fog where data is processed

5G are the most common network access employed, meaningfully. Architecture of the combat cloud‑fog con‑

especially with civil smart‑ phones [16] or drones [34, sists of the combat resource, fog computing and cloud

36]. With the expansion of smart cities, the public computing [41]. Combat resource is combat units which

WiFi hotspot is also a candidate. However, it is collect data and execute physical action, like radars or

noticeable that 4G seems to induce less la‑ tency than drones. They can communicate together. Networking de‑
WiFi in the case of video analytics [17]. Also, end devices
vices near the ield perform the fog computing. The com‑
and sensors may establish an ad hoc network to
puting tasks are distributed among them since networ-

communicate together and with edge computing de‑
king devices have their own duty and low capacities.
vices, without pre‑existing structure [17, 30]. The satel‑

Mission‑critical applications have strong requirements

lite network is an other option especially when the

re-gular network does not work, however the latency and distinct icities. We have seen they employ

can be problematic [16]. heterogeneous end devices, sensors, vehicles and au‑

tonomous agents, generating many data. All this data,
the unstable network and the strong latency requirement
3.2 Military make MEC a promising solution to deliver effective sup‑
port to agents. Moreover, we have seen diverse network

IoT for military is restricted because of unstable net‑ access used in these applications, which is consistent with

works, limited bandwidth, power‑limited devices and a

the variety of end devices. By integrating heterogeneous
highly dynamic environment [39]. Edge computing of‑
network access, MEC is all the more consistent in this type
fers the low‑latency and mobility required in the battle‑
ield [40]. of application. Finally, some applications employ vehi‑

cles to transport the edge server [16, 17], which stresses
3.2.1 Edge‑enabled military applications their high dynamic and need for resource mobility ma-
nagement.
The battle ield has a vast variety of heterogeneous sen‑

sors and devices that generate a lot of heterogeneous

data [41, 42]. To ease the instability of the network, edge
computing has the power to declutter all this
information by iltering, preprocessing and add meaning
to the mass of data. Singh et al. [40] introduce an
edge‑based system to monitor soldiers’ health, weapons
and location of those in command, the other soldiers and

themselves. In addi‑ tion to ilter, their framework
merges and attaches mea-ning to data to bring
situational awareness to agents on the ield and those in
command. Wang et al. [41] leverage fog computing to
compute and store the mass of data near the ield

and so provide real‑time responses. Moreover, they

use it to ilter and preprocess data to reduce

information sent to the cloud, sparing bandwidth. They
found that the latency is reduced to about 85% Fig. 2 – An example of an architecture of a MEC disaster management

when 300 tasks are requested. Castiglione et al. [42] use application

edge computing to authenticate agents with
their biometrics data when they access sensitive

4. RESOURCE ALLOCATION
material like weapons or vehicles, in addition to

monitoring their health. Lewis et al. [43] propose

tactical cloudlets to compute intensive tasks, like MEC resources are capacity limited, unlike the cloud, and

video and audio recognition and also ilter useless can even work on batteries. The resource allocation

data to lighten the application. scheme is so vital to manage these limited resources
64 © International Telecommunication Union, 2021