DRIVING BEHAVIOR TOWARDS VULNERABLE ROAD USERS : EFFECTS OF SAFETY COUNTERMEASURES AND DRIVING ASSISTANCE SYSTEMS

Abstract

The accident statistics all around the world show that the problem of the Vulnerable Road Users (VRUs) safety is relevant and much efforts are needed to improve the safety levels and decrease the accident events involving VRUs, as reported in the major government action safety plans. The higher probability of fatal event for pedestrians and cyclists in collision with vehicles leads to the definition of Vulnerable Road users. This definition implicitly considers the lack of protection that, together with the high difference in the physical mass, increases the risk of death in case of accident involving VRUs. However, as imaginable, reducing the mass difference is not possible, as it is an intrinsic characteristic of each road user. On the contrary, it is possible to improve the road configuration or the VRU facilities to induce proper driving behavior and reduce the probability of accident and, thus, the risk at which the VRUs are exposed to. Most of the literature is focused to analyze the behavior of the VRUs in the moment in which they interact with the drivers. However, it is highly important to understand how the drivers’ behavior change in the moment in which an interaction with the VRU occurs and how intervene to optimize this interaction and make it safer. Considering the safety conditions of VRUs, the caveats of literature and the actual body of knowledge the present research focused on the analysis of the driver - VRU interaction (pedestrian and cyclist) to provide further insights of this kind of road interaction. Furthermore, the research aimed also at assessing the effects of safety countermeasures at pedestrian crossing, cyclist crossroad and cyclist paths to detect the most effective driver/cyclist facility layouts and cross - section configuration that can decrease the risk at which VRU are exposed to. In addition, the assessment of driving behavior also under unexpected situation was analyzed to understand and verify the potential benefits of advanced warning systems aimed at timely alerting the driver about possible imminent collisions. The study of the driving behavioral models and how driver and VRU interact were the basis to achieve the research objectives. The foundations of the research are individuated in the theoretical framework of the driver behavior in response to a "threat" that is, in the present study, the presence of a pedestrian or a cyclist that should inevitably induce a change in the driving behavior. Several experiments have been carried out and specifically designed for the driver - pedestrian and driver - cyclist interaction. For the driver – pedestrian interaction, several countermeasures at the pedestrian crossing where assessed in the urban environment to identify the most effective one. Moreover, in addition to the improvement of the safety characteristics of the pedestrian crossing layout, dedicated experiments were designed to assess the effects of Pedestrian Protection Systems (PPSs) aimed at improving the pedestrian safety at the zebra crossing in urban, sub - urban and rural environments. The case of a pedestrian that crosses outside of the crosswalk (called jaywalking pedestrian) was also analyzed, for which the effectiveness of the PPS on helping the driver in this unexpected situation was assessed.As for the driver-pedestrian interaction at zebra crossings, several countermeasures at the bicyclist crossroads were evaluated in the urban environment, as well as the reorganization of the road cross - section to ensure the highest safety levels or the VRU during the driver overtaking maneuver of a cyclist in the rural environment. To achieve the objectives of the present research, the main tool used was the driving simulator. The experimental design followed several phases: a) literature review about the state of art of each specific topic investigated b) specific design of simulated scenarios c) simulated test on a significant diving sample d) robust statistical analyses on variables explaining the driver behavior. This approach ensured the objective comparison of the safety measures as well as the full control of the boundary conditions of the experimental road scenarios and the removal of the confounding factors that are often present in the field studies and that could affect the output of the analysis. Moreover, with the use of the driving simulator a huge amount of data is provided that can be used to produce significant statistical analyses and reliable outputs. The research findings provided further in-sights in the body of knowledge about Vulnerable Road Users. Specifically, the assessment of safety countermeasures at pedestrian and cyclist crossroad highlighted that drivers reacted differently based on the different interaction conditions with the VRU, highlighting different behavioral patterns. Specifically, aggressive drivers showed a tendency to delay the yielding maneuver to gain priority at crossroads, adopting more abrupt speed reductions. On the contrary, more cautious drivers tended to adopt smoother maneuver since they yielded to VRU from higher distances from the potential collision point. In absence of VRU, drivers didn’t perceive any interaction and tended to maintain unchanged the speed. These outcomes were consistent the behavioral model of Fuller (1984), confirming the reliability of the driving simulator for the study of the driving behavior. Furthermore, several safety treatments were investigated. The main findings highlighted that crosswalk with improved layouts increased the drivers yielding compliance (effect more evident for curb extension, -40% of non – yielding events). In addition, safety treatments that improved pedestrian visibility (curb extensions and parking restrictions) induced the driver start and complete the yielding maneuver farther (+21%) from the potential collision point. The model of the speed reduction time obtained through the survival analysis showed that for curb extension the yielding maneuver was smoother compared to the other treatments. The effects of the Pedestrian Protection System (PPS) were first evaluated under the presence of a pedestrian that crosses outside the crosswalk. Results showed that in presence of the warning system drivers reached lower speeds, decreasing the fatality probability in case of accident. Furthermore, a reduction of collisions was also observed (-63%). These positive effects were also evaluated under different road environments (urban, sub – urban and rural road). Results showed that PPS allowed the driver to advance the yielding maneuver (+9% farther from the pedestrian crossing) and to reach a lower minimum speed (-44%). At bicyclist crossroad, the less risky interaction was observed when the colored paved marking was present. In particular, the improved visibility of the crossroad gained the driver gaze attention; in presence of safety treatment a better yielding compliance and increased driving performances were observed. Specifically, lower speed and yielding maneuver farther from the potential collision point were recorded; this outcome highlighted better safety conditions for the cyclist that crosses the road. Finally, several cross – section configuration were evaluated to assess the driver overtaking maneuver of a cyclist in rural road. Results showed that for wider shoulder widths, higher lateral clearance was observed, accompanied also with decreased levels of influence of the cyclist on driver maneuver. This outcome is directly linked to the cyclist safety because higher lateral clearance affects the dynamic forces at which the VRU is exposed to during the overtaking maneuver.