A cross-study on video data gathering and microsimulation techniques to estimate pedestrian safety level in a confined space

Abstract

Nowadays, pedestrian safety is a growing problem: On the one hand pedestrian movement is increasing, on the other hand measures to improve walkers' safety are still scarce. During the years, various behavioural models have been proposed, which consider several parameters characterizing both the environment and the road users. Many studies define accident risk as a combination of three main factors, which are the accident occurrence probability, the vulnerability of the involved users and the economic and social damages due to the accident. The approaches spotted in literature about pedestrian safety assessment are various and they mainly deal with vehicle-pedestrian accidents. These studies have been developed in order to foresee and reduce possible losses in human lives. Also pedestrian-pedestrian encroachments can lead to dangerous consequences, but researches about them are still limited. This study aims to highlight the correspondence and/or the differences between different analysis methods applied to pedestrian field, focusing on situations where no interactions with vehicular traffic exist. This is one step of a wider research, which has been carried out on a confined space, selected because of easy replicability of its general geometrical features and of the effects on pedestrian trajectories induced by some local peculiarities of the facility. The study focuses on the evaluation of a real case, which has been monitored through the use of cameras and analysed via a dedicated tool. The achieved results have been then compared to the outputs obtained by reconstructing the same situation in a micro-simulation model. The first step-video acquisition and elaboration- A llowed to observe how people behave in the examined area and how the environment influences their trajectories, while the second phase permitted to understand if a microsimulation tool can reliably reply pedestrian movement in the analysed scenario and therefore provide surrogate safety values comparable to the ones obtained from real data. In the model, a simplified but still accurate environment has been set up: Homogeneous geometric features have been drawn and no obstacles have been considered. The modelled pedestrian flow is a bi-directional, 2400 ped/h flux, characterized by heterogeneity of agents: Both male and female adults. In order to be able to compare the data obtained by video footages and elaborated through an ad hoc tracking tool with outputs of the microsimulation model, from the whole flow on the ramp some intersecting pedestrians have been selected in both directions of walking. This cross-study of two different techniques has allowed to inspect the effects of the environment on pedestrian dynamics and to precautionary estimate the level of safety via calculation of surrogate safety parameters.