Connected and automated vehicle (CAV) has been recognized to have the great potential to improve traffic efficiency, safety as well as emission. Regarding the design of CAVs, the perception-reaction time (PRT) of human drivers is a crucial parameter to judge whether the CAV could detect and react more quickly and thus perform better than human. This study proposes two different methods to identify the PRT value based on microscopic vehicle trajectory data, the calibrated-based method and the duration-based one. The calibrated-based method utilizes a car-following model, intelligent driver model (IDM), to capture longitudinal driving behaviors and delineates the PRT as a model parameter. While the duration-based approach collects PRT by analyzing trajectory data of the car-following vehicle pair and extracts the PRT value manually. Two regression models are further employed to investigate the influence factors for these two types of PRTs. Results indicate the good performances of proposed methods and that the selected explanatory variables significantly affect the PRT. Findings of this study could provide useful information for the designs of CAVs in the future. Keywords: traffic safety, perception-reaction time, connected and automated vehicle, calibration, regression

CICTP