Modeling the Perception Reaction Time and Deceleration Level for Different Surface Conditions Using Machine Learning Techniques

Abstract

The ability to model the driver’s perception reaction time (PRT) and deceleration level is important for signal-timing design. The current state of practice considers PRT and the deceleration level deterministic and uses of constant values for them. The state of practice ignores the differences in PRT and deceleration level between individual drivers approaching the same intersection at the onset of yellow. The research presented in this paper uses data collected from two controlled field experiments on the Smart Road at the Virginia Tech Transportation Institute (VTTI) to model brake PRT and the deceleration level at the onset of a yellow indication for different roadway surface conditions. The paper uses many of the recent stat of art machine learning to train models that can be used to predict the brake PRT and deceleration level of approaching driver.