OBSERVATIONAL STUDY ON CHANGES IN DRIVER BEHAVIOR AFTER URBAN ROAD CORRIDOR RECONFIGURATION (ROAD DIET INTERVENTIONS)

Introduction and problem statement. Improving road safety is a top priority in the design of roads. The presence of different types of road users gives rise to the need of revising road layouts in order to ensure easy, fast, and safe mobility for all user categories. The concepts of Complete Streets and Road Diet have been already developed in literature, and guidelines are available to help engineers in taking effective design decision for new roads and in the re-design of existing roads as well. Moreover, studies have shown that by reducing the operating speed, the risk of collisions is significantly reduced too. The improvement of the road section layout, narrowing the lanes and separating them from cycle pathways, allows both reducing speed and improve safety. Road section modification are implemented worldwide but before/after studies to verify if the changes have led to the desired results are still insufficient. Research question. The aim of this thesis is to estimate the changes of speed and lateral distribution in the lane (i.e., road user’s behaviour) after the modification of a cross sectional characteristics of a street. The expected results consist in the reduction of speed and the decrease of lateral wondering. The analysis is focused on the most invasive user category (cars and small commercial vans). A 2.4 km segment of a collector road, via Nizza in Turin (North-West of Italy) has been considered as a case study. Re-construction works started in 2018; five out of nine investigated cross sections have been already converted. Data analysis was focused on those five sections. Methodology. Field surveys were firstly performed to collect data. Data collection was made through video shootings of about 15/20 minutes in each of the five sections and in two different days. Videos were acquired using a high-resolution camera, clamped to a carbon fibre telescopic rod, which was fixed to a tripod: this practical and affordable structure grants camera stability at 9-meter height, handleability and fast assemblage. In the laboratory, the recordings were analysed to extract vehicle speeds and trajectories. On the premise that road users have been divided into 3 categories. To avoid outliers and extract data samples of vehicles operating under free-flow conditions, some rules on the time distance between vehicles and on their manoeuvres were assumed. Pre processing of data as the Kolmogorov-Smirnov test and t-Student test were applied, respectively, to validate the assumption that the samples are normally distributed and to examine whether if different ones belong to the same population. Results and discussion. Tests confirmed that samples follow a Gaussian distribution. The mean comparison analysis indicated samples, belonging to the same section but to different time acquisitions, may not be brought together. Another error is given by the prospective view of the camera. This was identified clearly with the analysis of the Road Diet, thereby vehicles were aligned (less lateral wandering) as an effect of lane width reduction: distributions’ means depend on which side of the road the camera is positioned. Finally, an ANOVA was carried out to examine how big is the influence of road geometrical characteristics on the speed reduction and transversal deviation.