Aviation: Cyber Blackbox (Leading partner: TRT)

A Cyber Blackbox is a concept of aggregated monitoring of multiple avionics applications, possibly deployed on several Integrated Modular Avionics (IMA) heterogeneous computing devices, for safety, reliability and cybersecurity concerns. It would provide market advantages in terms of improved predictive maintenance, improved safety, introduction of adaptive cybersecurity and introduction of AI in the embedded avionics domain. Runtime monitoring events (such as runtime, data bandwidth or power consumption) from each monitored application are transmitted to the on-board Cyber Blackbox in charge of detection, analysis and proposition of mitigation of anomalous behaviour. It is implemented on heterogeneous hardware, including AI accelerators (Obj 1). The detection phase focuses on deviation from expected values for various monitored parameters: degraded performance, increased variation of execution time, additional data traffic, increased temperature, etc. The analysis phase uses AI algorithms for classification into e.g. safety errors or security threats (Obj 5). The mitigation phase also uses AI algorithms to propose actions such as restarting an application, ignoring a faulty sensor, etc. The pilot remains in the loop and can apply or ignore the proposed mitigation strategy, and that behaviour is monitored by the system to improve the analysis, via stick response-time and force-feedback measurement, wearable sensors (e.g. to track cardiovascular activity) or environmental sensors (e.g. camera and microphone). The Cyber Blackbox can integrate in a wider System of Systems, with intermittent connectivity (in-flight or at airports) with both cloud-based predictive maintenance centre and security operations centre collecting post-flight HUMS data and statistics (Obj 3). This concept will be applied in TEACHING to the monitoring of a Flight Management application, performing fusion of a number of heterogeneous navigation sensors and trajectory planning. Overall, this use case will exercise the support and management of heterogeneous, distributed and highly connected computing devices in the CPSoS, with a specific stress on the aspects related to energy-awareness, dependability and cyber-security of the autonomous applications within the TEACHING CPSoS (Obj 7). The demonstrator will serve as a basis for discussion with regulation authorities (Obj 8).