Artificial Engineering  –  AAA Process

History

The Artificial Engineer is a groundbreaking approach to the development and certification of safety-critical electronic flight control systems. It was developed by Aviotech based on fundamental research conducted at the University of Stuttgart, Germany.

Technical Basis

The Artificial Engineering (AE) approach for highly fault-tolerant electronic flight control systems is built upon two key innovations. These innovations enable a novel concept called Virtual Certification of applications.

Key Innovation 1: Platform Management Layer (PLAMA)

Unlike state-of-the-art approaches, PLAMA is implemented as a middleware layer.. It handles all communication, failure- and redundancy tasks necessary to operate an avionics system of distributed architecture including its associated redundant sensors and actuators in a highly fault tolerant way.  In doing so, PLAMA fully abstracts the complexity of the avionics system, providing a seamless interface for application development.

Key Innovation 2: Artificial Engineering – AAA Process

The Artificial Engineer is implemented as a comprehensive tool suite that captures avionics engineering knowledge using meta-models. Based on high-level system requirements, it automatically designs and instantiates all software components of the system platform, including:

  • PLAMA software,
  • Basic software (operating system and drivers) for all computing modules in the flight control system.

Furthermore, it defines and executes both system- and software-level tests while automatically generating all certification-relevant design and verification documents in compliance with avionics standards.

Virtual Certification of Applications (Cybernetics Tasks)

With PLAMA software in place, all system platform complexities become transparent to the application software. This means that applications operate in a virtually simplex environment, enabling a highly efficient certification process:

  1. All test-based verification activities for applications can be conducted in a simplified simulation environment.
  2. The Artificial Engineering process then transforms these tests for the real system platform, where they are automatically executed.

Maturity

The Artificial Engineer approach has already been successfully applied in multiple industrial technology programs in collaboration with well-known customers.