This paper presents a preliminary exploration of approaches to using experimental data for estimating the safety impact of advanced technology systems. The Crash Prevention Boundary (CPB) methodology is the basis for these new approaches. The CPB is an analytical technique to distinguish between driver performance that prevents a crash and performance that results in a crash. In this paper the CPB concept is used to describe the performance of an Adaptive Cruise Control (ACC) systems. Data from the Automotive Collision Avoidance System (ACAS) field operational test of an ACC system is used. This study explores a method to rate safety performance of ACC systems in two situations; where the host vehicle is overtaking a slower moving vehicle and where the host is following a lead-vehicle that is decelerating.
The paper presents an empirically based discussion of new computational procedures that can lead to improved estimates of the safety impact of driver assistance systems. The purpose of this paper is not to do a complete analysis of results from this test; but rather, to use a convenience-sample as a means of exploring new approaches to analyzing the data. The paper compares existing descriptions of safety boundaries with new approaches that are based on the CPB concept. Based on the ACC, it appears that these new approaches have the potential of improving the utility of such data for estimation of the safety impact of driver assistance systems.