In this paper, we present an example of how data from a recently collected set of naturalistic driving data can be used to improve the representation of driving conditions used in a comprehensive software simulation of headlighting performance. Simulations based on the detailed information about vehicles locations, speeds, and maneuvers provided by such a database appear to offer a valid way to quickly evaluate a large number of possible improvements in headlighting, including changes in static beam patterns as well as the dynamic adjustments in beam patterns used in adaptive frontlighting systems (AFS). Several software tools have been developed to simulate the overall performance of headlighting systems, including both glare effects and visibility. The CHESS system, developed by Ford Motor Company in the 1970s, was an early and prominent example of this approach (Bhise, Farber, Saunby, Troell, Walunas, & Bernstein, 1977). Such systems depend on several components, including representations of driver vision and target appearance. For present purposes, we will describe the CHESS system rather simply, in terms of two major components, which we will refer to as the vision model and the world model. By the vision model, we mean the way CHESS represents what drivers need to see (pedestrians and road edges), how they see it (based on the luminances of stimuli and the backgrounds for those stimuli, and including how glare from opposing headlamps affects those luminances), and the level of discomfort that they may feel because of glare from opposing headlamps. CHESS includes quantitative models of those aspects of vision, based primarily on laboratory data with some field checking. The importance of AFS for highly curved roads may be underestimated by the original version of CHESS, because the amount of travel on such roads is under represented in the CHESS world model. Speed as a function of road geometry should be more fully represented in the CHESS world model. The original version overestimates speed on shortradius curves. Overall, the approach used in CHESS appears to remain valid, and an appealing strategy would be to maintain the classic version of CHESS as a reference point, while still making major changes in its world model (dynamic details of driving, dynamic operation of lamps, etc.).
Software simulation of headlighting performance and naturalistic driving data
2009
7 Seiten, 4 Bilder, 5 Quellen
Aufsatz (Konferenz)
Englisch
Tema Archiv | 1976
|Modern Headlighting Requirements
SAE Technical Papers | 1934
|Kraftfahrwesen | 1984
|Automobile headlighting symposium
Engineering Index Backfile | 1925
Motorcycle Headlighting Research
NTIS | 1976
|