Safety of Unmanned Aircraft Systems Facing Multiple Breakdowns

Abstract

This work deals with data analysis issues in an aeronautics context by using a formal framework relying on activity recognition techniques which are applied to the certification and safety analysis processes of Unmanned Aircraft Systems in breakdown situations. In this paper, the behaviour of these systems is modelled, simulated and studied in case of multiple failures using a complex event processing language called chronicles to describe which combinations of events in time may lead to safety breaches, and a C++ chronicle recognition library is used to implement this method.