Abstract:
Pedestrians assess safe, available gaps in traffic flow and behave differently when crossing
midblock, making them vulnerable road users. The underlying factors of interaction
between the pedestrian and motor vehicle have a strong non-deterministic component,
making them uncomfortable and unsafe to cross. This paper explores various attributing
factors and interaction behaviors by modeling pedestrian gap acceptance and crossing
choice at uncontrolled mid-block road crossings. A field survey was conducted at three
uncontrolled mid-block crosswalk locations on major roads within the city of Shashemene.
Pedestrians crossing the road are videotaped in real traffic conditions, and the data is
extracted using a playback technique of AVS Video Editor Software. A non-linear model
based on an artificial neural network (ANN) has been used to build a model to assess the
impact of various factors on pedestrians' accepted gap size at uncontrolled midblock
crosswalks in mixed traffic. It was found that waiting time of pedestrians, vehicular arrival
rates, and type of gap have a significantly higher effect on the acceptance of traffic gaps. A
Binary Logistic (BL) regression model was also developed to look at various factors
affecting the probability of pedestrian gap acceptance. The finding showed that pedestrian
waiting time, crossing speed, and gap size have a much stronger impact on crossing choice.
Statistical analysis revealed that a gap size of 4.8sec and 6.8sec by 50% and 85% are
accepted by crossing pedestrians, respectively. The findings of this study will be useful to
increase the safety of crossing pedestrians by using it as reference for designing crossing
facilities and for controlling and managing the safety of pedestrian at crosswalk locations.
