You are on page 1of 15

AUTOMATION OF AVIONICS

INTEGRATION TEST
Hüseyin SAĞIRKAYA Neslihan Doğan AYDEMİR Ramazan ALTINSOY
Avionics and Electrical Engineering Avionics and Electrical Engineering Modeling and Simulation
Turkish Aerospace Turkish Aerospace
Simsoft Information Technologies
Ankara, TURKEY Ankara, TURKEY Ankara, TURKEY
hsagirkaya@tai.com.tr neslihan.doganaydemir@tai.com.tr ramazan.altinsoy@simsoft.com.tr

IEEE CIVEMSA 2021 – Paper ID


Content
1. Introduction
2. Test Environment
3. Simulator Properties
3.1. Test Definition
3.2. Execution
3.3. Tool Integration
3.4. Report and Analysis
4. Development Environment

2
1. Introduction
The system integration laboratory is built to accomplish system tests. Virtual environment is one of
the keywords of the laboratory to validate and to verify requirement, design and integrity.
There are 2 basic purposes to develop simulators. One of them is to test real devices, another one is
to replace devices. Using simulators decrease time, error and price.
ATLAS Test Suite is developed to fit above situations.

3
2. Test Environment
Tests run on simulation computers. Computers have communication cards such as MIL-STD-1553,
Arinc 429, Serial Port, CanBus and Ethernet also, simulators such as V/UHF Radio, IFF, ADC and
EGI also, platform sensors such as pressure and temperature etc.
Using simulators give an ability to switch real/simulator replacement piece by piece.

4
3. Simulator Properties
Simsoft has a product which is
called ATLAS that has some
abilities to prepare test cases and
execute them.
There are some modules in the
product that is shown in the figure.
In this paper, we focus on ATLAS
Test Suite. Test Suite has 4 layers;
definition, execution, requirement
links and report and analysis.

5
3.1. Test Definition
Using model-based approach brings some features.
The system does not need to be ready to prepare tests. System development and test case preparation
can continue in parallel.
It can be easily used by people with little technical knowledge as a result of using the tool. In
addition, user errors can be avoided.

6
3.2. Execution
Running tests can be operated platform independent.
Described tests have a run-and-forget concept.
As the number of tests increases linearly, the cost of running increases. For this reason, regression
tests are performed. However, this is not necessary with this tool. Full tests are not costly to run.

7
3.3. Tool Integration
ATLAS tool can be integrated with 3rd party software. This is very useful for requirements and
issues tracking. Examples of tools are Doors, Redmine, Jira and Gitlab.
Since the requirements are linked to the prepared test scenarios, it is possible to easily manage the
requirements.
After the test runs, reports are uploaded to the integrated tools.

8
Some Facts

Automated test preparation is


2 times faster than manual.

9
Some Facts

Automated test execution is


40 times faster than manual.

10
Some Facts

Automated test export is 17


times faster than manual.

11
3.4. Report and Analysis
As an example, report outputs can give the following metrics:
- How many test steps there are
- How many tests were run and faulty
- Causes of failed tests

As an example, analysis outputs can give the following metrics:


- Execution time
- Error rate
- How many requirements are connected

12
4. Development Environment
No coding knowledge is needed when creating test steps. There are some built-in steps like Dialog,
Wait, Send, Receive, Async, Code.
If uncommon behavior is desired to develop, ATLAS supports coding for model and test cases.

13
Thank you
Questions

14
Model-Based Development and Tools
• Allows a systems engineer to
create a model of the desired
system early in its life cycle
• Automatically generate code
and test cases. There are
already some tools for this
• Difficult to learn and time
consuming

15

You might also like