Automated measuring of cyclist – motor vehicle post encroachment time at signalized intersections
Conflicts between motor vehicle and cyclists at a signalized intersection were characterized in this study using an objective conflict indicator; post encroachment time (PET). A total of 384 conflict events for PET (0, 3] seconds between cyclists and vehicles were analyzed in this study. An automated video analysis technique was developed to measure the PET between cyclists and motor vehicles. The results of the conflict analysis showed that the average absolute error of PET between the frame count measurement (MFCM) and automated measurement (AM) methods was 0.12 s and the standard deviation was 0.10 s. The evaluation result showed that the coefficient of determination between the AM and MFCM methods was found to be 0.938 and there was a very good agreement in the PET classification of individual conflicts between the MFCM and AM methods. This study includes procedures to better interpret the conflict point of the motor vehicle and the cyclist in an automated manner (based on the geometry of the bounding box and direction of the travel), which appears to be a contribution for the analysis of cyclist – motor vehicle collisions.