Controllability of Active Steering System Hazards: From Standards to Driving Tests 2006-01-0600

When developing new automotive systems a great deal of the development effort is devoted to ensure a sufficient functional safety of the system. A question that arises during early risk analyses of such a system is that of the controllability of possible system hazards. While this question is answered in early stages very often using worst-case risk graphs, the question comes back later in a much more precise way: in case of active steering systems component failures would produce a deviation between desired and actual road wheel position, the deviation can be measured in terms of amplitude and/or time. The central question is how much deviation can be controlled by the driver? Note, that there will always be a certain, even small, deviation between desired and actual road wheel position since the steering systems controller contains feedback control algorithms aiming at minimising the regulation error but not actually making it disappear totally.

The contribution reviews the different notions of controllability used in safety standards such as MISRA Guidelines, IEC61508 [1], DIN V 19250 [6] and DS 00-55 [3]. The role of the operator/driver as a potential source of failure or as a safety measure is touched as well. Goal of this paper is to bridge the gap between safety standards and driving tests, recently applied during development of electronically controlled steering systems.