|

Cyclist Safety Behaviour Towards Stop Signs. A Study on Stop-Controlled Intersections Using Video Trajectory and Surrogate Measures of Safety

Author(s): Navarro, Saunier, Miranda-Moreno

Slidedeck Presentation:

4B Navarro

Abstract:

Background:

The treatment on the control type of the intersection in principle is justified from the traffic operation and safety points of view. In general, warrants justify the conversion of a MAS intersection into an AWS, or from MAS, AWS to traffic light signalized when traffic, geometry, and/or road safety issues are identified, and some basic conditions are met. However, those conditions do not consider cyclists or consider them as a pedestrian or as a vehicle, which does not reflect how they behave towards the other users of the road.

Aims:

The purpose of this research is to implement a statistical model to identify the behavior of cyclists at intersections controlled with stop signs and evaluate their safety risk with surrogate measures. More specifically, the objective of this work is to explore the bicyclist behavior to avoid users’ conflicts and speed reduction at the stop sign and their safety measures as post-encroachment time (PET). Additionally, the compliance of cyclists towards pedestrians at the stop sign is investigated.

Methods:

A high-resolution sample, video-based traffic collection system from 35 intersections is collected at five different neighborhoods in the city of Montreal. The information related to road users is extracted with the help of the tvaLib software, which is based on the open-source applied computer vision Traffic-Intelligence. Road users are handled as trajectories, and these trajectories are classified into three categories: pedestrian, cyclist, or motorized vehicle. Cyclist behavior as wavering, placing the foot on the floor were tagged, and their speed profiles with the average, minimum and maximum speed are determined for each user per approach with the extracted trajectories. Additionally, with the user’s trajectories, surrogate measures are obtained.

Results:

The evaluation of 4,923 cyclists in three different approaches conditions (without a stop sign, with a stop sign, or with stops in all the approaches) shows speed reduction between the cyclist without stop sign to the ones with the stop sign. However, the cyclists that are coming from an approach without a stop sign in a two-way stop sign intersection and the ones coming from an approach with a stop sign in a 4-way stop sign have a similar speed behavior. Also, surrogate measures are showing that the number of PET conflicts persists, the median of the values is slightly moved towards a safer number; however, this result is not significant.

Discussion:

The PET analysis shows that the safety risk that the cyclist would put on the other road users does not have a difference between the different evaluated scenarios. Cyclist on average not following the rule of stopping at the stop sign as the rules of the road indicated. Yet, bicyclists are taking precautions to avoid conflicts with pedestrians and motorized users as is shown in this research.

Conclusions:

It is found that having stop-signs in the approach reduces the predicted mean cyclist speed around 1.0 km/h. Even though it is a small reduction, it is significant for the different speed measures. Also, it was found that less than 2% of the cyclist do a complete stop placing their foot in the ground. Nonetheless, cyclists compensate for the non-stop compliance with a wavering movement towards pedestrians to avoid conflict and to continue with the movement.
Nevertheless, this research in the city of Montreal is showing that cyclists could use stop signs as yield signs, improving their comfort and travel time while commuting, without increasing the probability of a collision at the intersection level.