Abstract
In first embodiments of automated driving systems the customer acceptance will depend on well-defined cooperation between the driver and the automated vehicle. Design rules for supervised driving automation have to be identified and confirmed. The capability to establish such guidelines depends on understanding and analysis of quantitative and qualitative traffic factors in a real environment. The selected automation system under review is called Traffic Jam Assistance (TJA). Based on data of a naturalistic driving study in traffic congestions in Europe the investigation builds on the method of video data coding and its statistical analysis. Analysis of video and sensor data revealed potential indicators for relevant traffic manoeuvres such as cut-in or cut-out. These indicators serve to support the future design of a warning taxonomy defining when the driver either needs to regain situation awareness or a transition from automated to manual driving has to be initiated.
KEYWORDS –Traffic Jam Assist, Naturalistic Driving Study, Lane Change Analysis, Transition Taxonomy, Video Labelling