Revising the DETECT Seeing Distance Model

This paper describes the algorithms that are being incorporated into a revised version of the DETECT seeing distance model. The revised model is being written in QuickBASIC for IBM-compatible personal computers. DETECT calculates the distances at which a driver can see various objects on the road at night as illuminated by the headlamp system specified by the user. The revised algorithms are based on professor Blackwell's recent contrast sensitivity research. They include new formulations for calculating contrast thresholds and for taking driver age, target size, background luminance and individual differences into account. The revised expression for disability glare also has a correction for driver age. The seeing distances calculated using the old and revised versions are generally in close correspondence. However, at low illumination levels, the new algorithm predicts seeing distances that are as much as 12% greater than the original version. These differences can be traced to differences in between the old and revised contrast threshold functions. The age and variability factors in the new algorithm have a substantial impact on seeing distances. Against a low beam glare source at 300 feet, the seeing distance for an average 20-year old is 210 feet while for a 15th percentile 70-year old it is 92 feet.