On 4th of November, SYSGO participated in the High Integrity Software Conference (HIS 2019) in Bristol, UK. The talks were accompanied by an exhibition, where SYSGO and their partners were showcasing PikeOS avionics and automotive solutions.
The visitors and exhibitors of the HIS are a community of safety and security specialists, where all know each other and it felt like a meeting of friends. At the booths of Adacore, ANSYS, RAPITA SYSTEMS, rti, VECTOR and SYSGO itself, the visitor could get a good overview about the PikeOS ecosystem and certification solutions in terms of safety and security.
The speeches were mainly influenced by upcoming changes regarding the development and certification of safety critical systems that are being driven by innovation. The keynote, held by George Romanski (FAA), explained the need for cost effective certification methodologies in order to get more safety enhancing components inside avionic systems.
The FAA is exploring a new concept that is called “overarching properties”. This approach puts the idea of the” intended behaviour” and “assurance cases” of a system in the foreground. The work on the idea of overarching properties is still in its early stages and some issues need to be clarified, such as how to prove the completeness of the description of the intended behaviour and the assurance cases.
In the future, the FAA plans to extract the essence of existing safety standards (such as DO-178, ISO 262626, etc.) and channel it into a higher-level description that is called the “Abstraction Level “. Isolation, abstraction and certification standards is clearly within the domain of SYSGO and its proven PikeOS RTOS.
There were also the inevitable warning voices: There is a trade-off between high certification need and quick innovation as more innovation needs more software use in airplanes.
Prof. Les Hatton stated that software should not take primacy in avionic systems and substantiated his view by several examples, including the most recent Boeing 737 MAX incidents. This was in direct contrast to Mr. Romanski’s position where safety can be enhanced significantlyby using software in order to make the handling of different types of airplanes more consistent and easier to fly by pilots.
What do you think? Should software be used in order to reduce the complexity of airplanes and make flight more accessible and uniform to pilots, even when that means masking distinct flight characteristics? Is it too dangerous to drive this kind of innovation?
Please send us your opinion!