Aerospace and Electronic Systems - March 2019 - 14

The Adaptive Avionics Platform
Table 1.

Average Times for the Steps of Self-Organization for Different Topology Sizes (10 Runs Each)
Number of Modules
(#CPM þ #IOM þ #PERIPHERAL)

Time [s]
2þ1þ0

2þ1þ1

2þ1þ2

2þ3þ9

Identification and topology discovery

61.2

52.0

58.7

52.7

Peripheral discovery

18.3

18.9

19.8

19.4

Application distribution and initialization

4.2

4.1

3.8

4.5

Routing

61.9

61.5

60.1

59.2

Self-verification

9.2

9.5

10.8

11.8

Entering the standard phase

2.4

2.8

2.5

2.7

Total

157.3

148.8

155.7

150.3

controlled locally or centrally, depending on the
flight phase.
Audio (AU) comprises the crew communication with
handsets, live audio announcements, and the playback
of recorded announcements [30]. This is an example
of an application that requires high bandwidths, multicast capabilities and a low 10 ms delay [29].
Temperature control (TC) comprises the reading
and displaying of the temperature value of the TS.
It is an example of the integration of a legacy
peripheral.
The FAP is the command center of the crew.
It depicts PSU statuses and remote control PSUs, IL,
and AU. Moreover, it is used to set mission modes.

VERIFICATION OF THE ADAPTIVE PLATFORM CONCEPT
Using the ACMS components, self-configuration tests,
automated topology verification, and fault-tolerance tests
were carried out to verify the AAP concept.

SELF-CONFIGURATION TESTS
The proper functioning of the AAP mechanisms for selfconfiguration was tested by putting together several of
the meaningful topologies. Each topology was physically
built, and then the self-configuration process on the
CORG was started. Furthermore, it was checked if
the self-configuration phase was successfully completed
and if all applications and peripherals (including legacy
and smart peripherals) worked correctly. For all the tested
architectures, the organization and standard phase worked
as expected. The self-configuration process was completed
14

in a mean time of 153 s. The times for the steps of the selfconfiguration process for different topologies are listed in
Table 1. No significant increase of the self-configuration
time for increasing topology sizes can be stated (at least
for the variations possible with the ACMS components).
The dominant steps are module identification and discovery, peripheral detection, and routing.

AUTOMATED TOPOLOGY VERIFICATION
Automated topology verification was tested by building
several ACMS topologies, inserting errors, running selforganization, and verifying the correct detection of the
errors. Besides the real ACMS architectures, the ground
truth architectures were modeled using the Open Avionics
Architecture Model (OAAM) [31]. The ground truth was
provided to the CORG and is compared automatically at
the end of the organization phase. Four classes of topological errors were tested: 1) Missing and extra modules,
2) missing and extra peripherals, 3) swapped connections,
and 4) missing or extra partitions. 100% of the deviations
were detected. Figure 8 depicts three artificially introduced
topological errors and the verification output: a) An IOM
was expected, but is missing in the real architecture.
b) A peripheral and its partition were expected, but not
discovered in reality. (c) The interfaces of PSU (smart) and
TC (legacy) on IOM1-1 were swapped. The CORG generates the topology graphs automatically and shows them
with problems highlighted if the topology verification fails.

FAULT-TOLERANCE TESTS
Fault-tolerance verification tests were carried out during
the standard phase in three categories: 1) Network faults,
2) IOM faults, and 3) CPM malfunction. Faults were

IEEE A&E SYSTEMS MAGAZINE

MARCH 2019

Aerospace and Electronic Systems - March 2019

Table of Contents for the Digital Edition of Aerospace and Electronic Systems - March 2019