By studying drivers’ behavior, driving simulation is used in automotive industry for designing and testing new driving aid systems. In order to have a behavior similar as much as possible to the one observed in real conditions, driver has to be provided with visual, audio or kinesthesic cues as well as inertial cues. A “one to one” motion rendering is usually not possible due to physical limitations of dynamic driving simulators, so a so-called “motion cueing algorithm” is used to transform virtual vehicle trajectory into admissible simulator trajectory. Our knowledge of the human motion perception is currently incomplete. A way to improve motion rendering is to increase driving simulators physical abilities. “High performance” driving simulators thus obtained can provide inertial cues nearer to those in real conditions, but they need large simulation rooms and complex operational facilities. The second manner to improve motion rendering is to develop new motion cueing algorithms, and this is what is proposed in this paper. In the framework of a partnership between Arts & Me´tiers ParisTech and Renault, a new dynamic simulator called SAM has been built. This simulator is equipped with traditional hexapod motion-platform, nevertheless it is using an innovative motion cueing algorithm. In this paper, an overview of existing motion cueing algorithms will be presented, especially their limitations and the relevance of a predictive algorithm. Finally, an experiment will be also presented for comparison of the different cueing algorithms.

This content is only available via PDF.
You do not currently have access to this content.