The multitude of activities in the area of vehicle dynamics covering the sector of calculation as well as measurement technique can lead to the wrong conclusion, namely that today we would be able to measure and describe vehicle dynamics objectively. In truth the whole field is of such a huge size that we have just been able to measure and describe vehicle dynamics objectively.

A special consideration of the control-loop environment-driver-vehicle gives insight into the real relations.

The possibility of bad processing of information by the driver is much higher than for the vehicle and the part of the still improvable vehicle behaviour that will be utilized by the driver is very small. This leads to the realization that, by measures to the vehicle, only a few additional accidents can be avoided.

The small scope of significance of today's known test methods and the lack of knowledge about the relation between test results and safety lead to the conclusion that exhaustive statistical investigations of driver behaviour prior to accidents have to be undertaken and that, after accomplishing this, handling rules directed toward the specific objectives will be useful.