Airborne Separation Assurance Systems: towards a work programme to prove safety

With Full Delegation Airborne Separation Assurance System (ASAS), separation control would be delegated to the (properly equipped) aircraft, i.e. aircraft pilots are responsible for aircraft separation. The aim is to try to identify a tangible work programme-rational and evidence based, and within the compass of known techniques-that would prove safety. The task here is to create a framework in which to integrate these existing building blocks with results from additional work developed from well-specified experiments. Reasons for retaining the existing separation minima in an ASAS system are put forward. For the current system, comparatively large proportions of the Air Traffic Services risk budget should be allocated to 'Reasonable Intent' risk (effectively 'right place on wrong flight path'). The key argument here is that mid-air collision in an ASAS environment will predominantly arise from this type of risk. The use of probabilistic risk assessment, which requires the probabilities of safety-critical events to be estimated for 'human components' (human reliability analysis), is reviewed. The danger is the creation of 'over-elaborate' models-ones whose parameters cannot be reliably estimated from the data likely to be obtainable. A simple model that can be soundly based on available data is proposed. (C) 2004 Elsevier Ltd. All rights reserved.