Automated testing and specification mining for certification of avionics software systems

航空电子软件系统是嵌入在飞机硬件中的安全关键软件，用于控制和监控飞机。无人机（UAV）是一种没有飞行员的航空电子系统。无人机可以使用名为Autopilot的软件自主飞行。 航空软件行业遵守最高的安全控制标准之一，他们的嵌入式软件系统，如自动驾驶仪，在进入商业市场之前，要经过一系列严格的标准检查。在这个项目中，我们与MicroPilot合作，MicroPilot是为商用无人机构建自动驾驶系统的世界领导者，通过提供“自动化”，“可追溯性”和“风险意识”来改善他们目前的测试和调试实践，这些都是标准中最重要的要求。为了实现上述目标，一套技术，从“挖掘软件仓库”，“模型驱动工程”，和“自动化测试数据生成”将被用于本研究项目的背景下。该项目的拟议工具和流程计划集成到公司现有的框架和流程中，以帮助他们获得认证并提供更高质量的系统。虽然目前的项目将侧重于改进MicroPilot目前的软件开发和验证实践，但在加拿大建造安全关键嵌入式系统的其他公司也可以从该项目的报告中受益。

Avionics software systems are safety-critical software embedded in an aircraft hardware that control and monitor the aircraft. The Unmanned Aerial Vehicle (UAV) is an avionics system with no pilot on board. A UAV can fly autonomously using a software called Autopilot. Aviation software industry observes one of the highest standards of safety control, where their embedded software systems, such as an Autopilot, go through a set of rigorous standard checks, before entering to the commercial market. In this project, we collaborate with MicroPilot, a world-leader in building autopilot systems for commercial UAVs, to improve their current practice of testing and debugging by providing "automation", "traceability", and "risk-awareness", which are among the most significant requirements of the standards. To realize the above objectives, a set of techniques from "mining software repositories", "model-driven engineering", and "automated test data generation" will be used in the context of this research project. The proposed tools and processes of this project are planned to be integrated in the existing frameworks and processes of the company to help them to get certified and deliver higher quality systems. While the current project will focus on improving MicroPilot's current practice of software development and verification, other companies that are building safety-critical embedded systems in Canada could also benefit from this project's report.