Year: 2005
Type: book-chapter
Abstract: The goal of our work is to develop a methodology for automatically predicting potential pilot errors during the design of mode-based systems for modern cockpits. The main ingredient of this methodology is a cognitive model of pilot behaviour allowing to predict misconceptions in the pilot's mental procedure model based on a cognitive process called “learned carelessness”. In a first stage we use this model inside a simulation platform to automatically degrade a normative procedure model in a psychological plausible way. In a second stage we apply formal verification techniques to exhaustively analyse what human errors can be caused by the degraded procedure. In this text we describe a case study conducted at the Lufthansa Flight Training Centre where we analysed a Piper Cheyenne autopilot. 1 Human Error Analysis in Early Stages of Safety Critical System Design It is often cited that 60-80 % of commercial aircraft accidents are caused by the flight crew. What is the reason of this high percentage? The Federal Aviation Administration (FAA) conducted a study to investigate the interaction between flight crews and modern cockpit systems (like flight management and autopilot (AP) systems). One major concern highlighted in the study is a general lack of mode awareness that exists independent from the manufacturer, operator, or airplane type (Lyall & Wilson, 1997). A mode may be understood as a system state in which it delivers
Authors: Andreas Lüdtke, Claus Möbus
Institution: Carl von Ossietzky Universität Oldenburg
Language: English
Subfield: Human Factors and Ergonomics
Field: Social Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W2511934482