Handling Complex System Architectures with a DO-178C/DO-331 Process-Oriented Build Tool

Abstract

Software development in safety-critical systems is invariably accompanied with extensive documentations, strict methodologies and verification activities. While software vendors will provide the necessary software tools and tool qualification artifacts, the details on how each tool component is interlinked in development process are usually a part of the intellectual property of large aerospace companies and not publicly accessible. This poses a market entry barrier for startups and small/medium enterprises, whose numbers have grown, especially in the areas of electrical aviation as well as unmanned aerial vehicles (UAVs) and electric vertical take-off and landing (eVTOL) systems.The process-oriented build tool presented in this paper is aiming to address this problem by providing an exemplary toolchain setup for a DO-331 compliant software development process. Based on MathWorks' MATLAB and Simulink products, the tool provides a development environment with predefined model templates, block libraries, and configuration settings as well as jobs for executing process-relevant tasks, like automatic code generation or static model analysis. By doing so, the tool ensures consistency of model artifacts created by developers across teams and also compatibility with downstream tools used for verification and validation on model and code level. Artifacts from each process step are stored within the tool so that full bidirectional traceability can be ensured.While the tool has been used in the development of flight control applications in the past, its capabilities are currently improved based on lessons learned from these projects and furthermore, extended to new use-cases.This paper will discuss two tool improvements: handling of dependencies of distributed software modules and tool artifact ownership, which are made to handle complex software-development project consisting of multiple software components developed by a distributed team. To demonstrate the improvements, the development of a distributed battery control software used in a smart-battery concept for an electrically powered aircraft is presented. This software is comprised of multiple software modules representing a battery master controller as well as multiple slave controllers.