Flights are a special experience for many people: walking to the entrance of the plane, welcoming the crew, and taking off to the desired destination. On the other hand, there are many who fear the experience due to possible accidents. This concern increases as aircrafts become more sophisticated and flight systems become more software-based.
To alleviate these fears (but actually to mitigate accidents) specific safety standards were created. Just like there is a dedicated and specific standard for the development of the mechanics and hardware of the aircraft, there is a mandatory standard for developing aviation software – DO178.
Since the introduction of the DO-178 standard in the 1980s, no accidents related to software failures have been officially documented. In addition to these excellent statistics, software testing is constantly improving with the technological advances in flight computers and software.
The main challenge is the detection of defects as early as possible in the development stages and of course to reach zero defects impacting the customer and especially defects that are reported as “critical”.
This is where the automation of software testing in aviation comes into play. As flight technology advances and code becomes more complex, the importance of automated testing increases.
For example, Qualitest Aerospace Division analyzed the data from projects with very similar software and compared the number of defects discovered with manual testing versus automatic testing.
The results were indisputable in all samples – automated tests found significantly more defects than with manual tests.
The explanation for this is simple. Tests that are based on a visual validation of the tester may be biased due to the tester’s errors, as opposed to machine tests, which depend solely on the quality of the code that runs the automation.
The benefits of automated testing in aviation do not stop at the quality of testing. With automated testing we also gain savings in running time, manpower and of course significant cost efficiency.
When automated testing becomes a standard practice in aviation testing, the next step is to assimilate AI capabilities in the development process.
AI-based testing, especially in image processing, which is the biggest current technical challenge, will lead to a further increase in the number of defects found in the early stages of development and as a result to higher software reliability.
