Hazards in advising autonomy: Incorporating hazard modelling with system dynamics into the aerospace safety assessment process for UAS

摘要

This paper describes the further continuation of an investigation to identify and develop tools for the identification and management of hazards likely to arise with the quality and behavioural aspects in and resulting from automatic advice - such as might arise with an automated system advisory function facilitating critical decision-making with an autonomous vehicle. An example of a representative critical advisory function is identified in that supporting a necessary “Sense & Avoid” capability, as embodied within a airborne autonomous system. In consideration then of how might a model driven approach, combining physical and dynamical models, statistical data and belief be combined to aid system evaluation, work has so far been undertaken to investigate the nature of suitable models to provide representations of the control structure and system dynamics. Whilst the system engineering methods are to be generic, the context of “Sense & Avoid” provides a relevant framework within which to pose a “toy-problem” with complex behaviour, against which to judge the methods and models.