IEEE - Aerospace and Electronic Systems - December 2023 - 24

AI-Powered Edge-Cloud Continuum for In-Flight Entertainment and Connectivity
Figure 8.
Inner cabin aeronautical software defined radio network.
section " Interference Detection and Localization. " Information
about detected radio jammers, provided via
MQTT, can be accessed by the crew recommendation system
to trigger individual tasks to the cabin crew members
to identify the jammer and in the case of a jamming passenger
ask him to turn the device off.
The triggering of tasks is carried out through graphical
user interfaces (GUIs) for the cabin crew. Main GUI are
accessible for all crew members at central locations, such as
one notebook-sized touch screen per galley in the aircraft.
While the main GUIs are not assigned to dedicated crew
members, each crew member carries their own smartwatch,
which is directly connected to the inference module via
wireless local area network (LAN). General management
tasks ofcabin operations and dedicated interactions with the
crew recommendation system can be performed by the crew
members via the large screen main GUIs, whereas the smartwatches
provide a direct interface ofthe system to each crew
member without affecting the cabin crew in their tasks during
carrying the devices. Smartwatches are ofspecial significance
to the system for triggering individual tasks to the
crew members and for triggering time-critical emergency
tasks. Various notification functions of the smartwatches,
such as screen-blinking, acoustic alerts, and vibration alerts,
facilitate adequate notification of the crew members regarding
tasks and information.
For decision-making regarding the next tasks in cabin
operation, the inference module carries a static knowledge
database. It comprises a general dataset of cabin operation
tasks to be carried out during each flight. Besides normal
operation tasks, the database also lists emergency tasks,
which can be triggered by the crew members themselves or
external systems like the jammer detection system. Furthermore,
the knowledge database holds templates of catering
items, which are available at the airline. For adapting the
static task structure of normal operation tasks and general
catering tasks to the individual upcoming flight, a flight profile
is loaded to the system previous to each flight. The flight
profile may be part of the pre-flight briefing of the aircraft
crew and it is uploaded to the system by the cabin crew
directly after crew-boarding. The flight profile specifies
parameters like the number of passengers, composition of
passenger categories (first class, business, economy), flight
24
duration with estimated timing offlight phases, and information
about the available catering categories.
INTERFERENCE DETECTION AND LOCALIZATION
The AI-supported in-flight edge-cloud network
approach includes also surveillance, which involves
detecting and localizing radio-based sources of interference.
This approach categorizes interference as intentional
jammers and nonintentional interference (such as
out-of-band devices). Further approaches are developed
with regard to spoofing, i.e., the deliberate identityrelated
falsification of radio signals and their content
[16], [17]. To achieve the overall objective of monitoring
the entire state of the edge-cloud network based on
continuous analysis of radio-based in-flight components
for the aircraft cabin, a hybrid state monitoring system
is being developed. This system utilizes a radio-based
aeronautical software defined radio network (aSDRN).
The aSDRN consists of software defined radios
(SDRs) distributed in the aircraft cabin, which are geometrically
arranged in such a way that they can provide radio
coverage of the entire cabin interior (see Figure 8).
The SDRs perform two different tasks. First, they detect
the sources of interference described above, and second,
they localize them. In general, it is possible to configure and
control the SDRs with special software so that they scan and
analyze certain frequency ranges for specific effects. This is
done with a combination of real-time spectrum analysis in
conjunction with historical measurements, for example, to
prove the coexistence between a wireless narrowband sensor
network and a broadband wireless LAN system. Both are
visible in real-time mode in the spectrogram as well as in the
frequency spectrum. With the evaluation of this real-time
spectrum, which runs as a " waterfall " in time, one can capture
bandwidth-related spectral components of all signals
occurring in the cabin area. The evaluation of these spectral
components, including pattern recognition for interferers
and jammers, will ultimately only be possible with an AI
approach. The limits for the use ofSDRs are set on the hardware
side by parameters such as the frequency range, bandwidth,
performance of the analog/digital and digital/analog
converters, and clocks used. In this regard, potential
IEEE A&E SYSTEMS MAGAZINE
DECEMBER 2023
IEEE - Aerospace and Electronic Systems - December 2023

Table of Contents for the Digital Edition of IEEE - Aerospace and Electronic Systems - December 2023
