Estimating potential conflicts between right-turn-on-red vehicles and pedestrians at crosswalks

The conflict between vehicles and pedestrians at crosswalks, defined as a potential collision that could lead to actual traffic accidents, is an important measure indicating the level of pedestrian safety. Past research, with advanced statistical methods, gathered and calibrated data from historical accidents to produce this important policy indicator. To perform such statistical analysis, however, requires a considerable amount of samples collected under controlled conditions. Putting aside the difficulty of creating a controlled condition, it is difficult to apply the model in different locations with various physical conditions. These are critical limitations in practice. This study aims to overcome these issues and develops an analytical model for estimating the conflict between vehicles and pedestrians. Since the proposed model considers pedestrian behaviour at crosswalks rather than actual collision samples, only a small number of field data are required. It appears that our model can be replicated to different times and places with a bias that is statistically insignificant.