AdaCore has announced that MASC (MHI Aerospace Systems Corporation), a member of the Mitsubishi Heavy Industries Group, has selected the QGen toolset to develop the software for the TQA (Throttle Quadrant Assembly) system.
This avionics research project is being conducted to meet the Level C objectives in the DO-178C safety standard for airborne software and its DO-331 supplement on Model-Based Development and Verification. The use of a qualified code generator can help save significant effort in developing and verifying the software, and the future availability of qualification material from AdaCore factored strongly in MASC’s decision to choose QGen. The QGen code generator can be qualified at the highest Tool Qualification Level, TQL-1 (equivalent to a development tool in DO-178B).
“The QGen product is specifically targeted to model-based development in safety-critical control systems,” said Juan-Carlos Bernedo, QGen Product Manager at AdaCore. “However, model-based development raises some important questions: how to verify the model’s safety properties, such as freedom from run-time errors, and how to know that the properties are preserved in the generated code. QGen and its supporting TQL-1 qualification material help to answer these questions, and we’re pleased that MASC’s Throttle Quadrant Assembly project will be taking advantage of these benefits.”
“We chose QGen as our Auto Code Generator because we believe TQL-1 qualification will reduce the effort of our verification activities that comply with DO-331,” said Hiroyuki Kakamu, General Manager of MASC. “As part of the TQA project we discussed DO-331 compliance with AdaCore’s engineers. Based on these discussions we developed a Model-Based Design process that complies with DO-331 based on the use of QGen.”
QGen’s qualifiable and customisable code generator processes a safe subset of Simulink and Stateflow models and generates optimised source code in the safety-oriented programming languages SPARK (a formally analysable Ada subset) and MISRA C. QGen incorporates static model verification, PIL (processor-in-the-loop) testing on a real target or through emulation on the host robust, and powerful model-level debugging support for back-to-back testing between simulation and target execution.