SCENARIO-BASED ANALYSIS OF FORWARD CRASH WARNING SYSTEM PERFORMANCE IN NATURALISTIC DRIVING

This paper uses empirical data from 925 miles of driving with two early prototype FCW systems to illustrate the utility of a scenario-based analysis of performance. The systems are immature and the experimental scope is limited, but the data is used to identify driving scenarios associated with crash alerts. The scenarios are grouped according to the primary elements of the conflict, which also correspond to FCW system components of alert timing, path prediction, road geometry prediction, and sensor errors. The relative frequency of alerts in these scenarios therefore indicates the relative value of improving the FCW system components. The influence of road type and driver actions is also explored as factors in the occurrence of alerts in the different scenarios. The scenario-based approach may be a useful tool in the design and evaluation of FCW systems.