This presentation demonstrates various applications of the SAFE methodology (SAFE: Situative Analyse von Fahraufgaben) for the driving task analysis and driver behaviour requirement assessment. A framework for task analysis is derived from a model of the drivers´ information processing. The first step of the procedure is to divide a given driving task (e.g. left turning in a signalised intersection of a certain type) into subtasks (about 30 for the example). These subtasks are appointed to defined stretches of the road and the time structure of the subtasks is determined. For each subtask an analysis format is used, that organizes different requirements into perception, expectation, judgement, memory, decision and driver action. In a next step, typical driver errors are appointed to the subtasks. Afterwards the information gathered in the analysis procedure is compressed to ratings of the complexity and risk included in each subtask. After the crucial subtasks of a driving task have been determined, a list of those requirements that led to that rating can be compiled. First we have used the results of SAFE to find the crucial subtasks that may be supported by driver assistance systems. A list of those requirements can be compiled for each subtask that led to the rating of high complexity or high risk. These lists accurately tell the system designer, which requirements need assistance and when and where system functions should give information to the driver or intervene. SAFE has also been used to compare the task structure between driving with vs. without a defined driver assistance system (ADAS). This comparison can already be accomplished in an early development phase of a system. It shows which variations in task complexity and risk are to be expected by the introduction of an assistance systems and thus gives valuable feedback into the design process. This idea is demonstrated using an example of a system designed to help the driver in the approach to signalized junctions. Another utilization of SAFE lies in the fact that driver errors now can be defined as deviations from a well defined correct behaviour. We have conducted a field study with 62 subjects and compared their actual driving performance to the SAFE standards. Some principle results concerning the validity of SAFE as a predictive tool for driver error are shown. Finally, a number of further applications of SAFE, e.g. in driver training and traffic engineering, are presented. (Author/publisher) This publication may be accessed by Internet users at: http://www.ictct.org/workshop.php?workshop_nr=25
Abstract