Advanced Driver Assistance Systems for Active Safety of Modern Tram

Abstract

Mobility in smart cities is also becoming smart, promoting on the one hand transport modes based on zero-emission electrical technologies and on the other providing vehicles with technological solutions that support the drivers in driving operations. The winning key for mobility in the cities of the near future, therefore, are electrified transport solutions with assisted driving: the tram is the ideal candidate for this revolution of green and safe mobility. Only in 2019, road accidents in the European Union summed up to 22.660 fatalities and 120.000 people seriously injured. However, the trend of accidents is decreasing every year and looking to the past, the decrease coincides with the market penetration of Advanced Driver Assistance Systems (ADAS). Nowadays, almost all manufacturers offer cars with level 3 of driving automation (SAE 3 levels) and many prototypes of self-driving vehicles are developing and circulating on our roads. From a medium-term perspective, all vehicles running in the smart city will be equipped with autonomous driving technology or advanced driving aids, including public transport vehicles, such as buses. It is clear, therefore, that even for trams, the time is ripe to accommodate driving support systems. Just the tram is

Abstract and activity form of Catello Di Palma – XXXIV – I-II-III year

experiencing a new spring today, due to its green nature that makes it a candidate for sustainable city mobility, and it will be the next candidate to host ADAS technologies. Trams are the combination of two worlds: railways and the road environment. Infect, unlike the other rail transport systems, use the same road infrastructure as cars, motorbikes, bikes, and pedestrians and will soon find themselves interacting with vehicles with increasingly higher autonomous driving levels. It is clear, therefore, that even for trams, the time is ripe to accommodate driving support systems. From a technological point of view, tram manufacturers are moving towards the use and integration of automotive solutions, already available with a high level of maturity and reliability, based on Sensor Fusion and Perception Platforms. However, due to the very different stopping times, a careful analysis of the tram braking dynamics to identify the most suitable technological solution is necessary. The purpose of my thesis is to provide a contribution to the implementation of a driving assistance system designed for the tram, here called T-ADAS, by developing a model for Estimation of the Distance to Collision (DCE) fitted to the tram. In the thesis, the key elements of technological porting and the choice of functionalities for a T-ADAS system according to the Degrees of Automation (GoA) are initially investigated and defined. The degrees/levels of automation in automotive and railway systems are presented and compared to each other. Then, according to the implication level of a remote sensing system in each tram driving task, new GoAs for tram are proposed. These systems are designed to help the tram driver cope with potential hazards by having defensive driving. Therefore, the proposed GoAs correlated to ADAS will be useful to understand how this automation acts the action of the tram driver, and the safety of the entire transport system. The problem of braking for the tram is then analyzed to define quantitative requirements for the design and validation of the Forward Collision Earning (FCW) and Autonomous Emergency Braking (AEB) systems that takes into account the very different dynamics from the car. Those technologies are today widespread in the automotive sector: in fact, vehicle collision warning systems have been studied by

many researchers, and many approaches related to technologies and problem formulation has been developed and a lot of commercial solutions today are available. Among the ADAS, FCW-AEB systems represents the one with the highest percentage of crash avoidance effectiveness: FCW alone, low-speed AEB, and FCW with AEB reduced rear-end striking crash involvement rates by 27%, 43%, and 50%, respectively. For this purpose, the proposed DCE model, which uses tram and tramway data (mass, power, grip, slope, radius of curvature) and real-time data (speed, weather conditions, and GPS) will allow the calculation of the safety braking distance. The main difference between the model proposed and those of the literature and/or of the commercial ones, is mainly related to the vehicle dynamic. Therefore, in the proposed approach, instead of considering only the deceleration and speed data to calculate the stopping distance, we consider many other important factors that tune this calculation to the real one. The robustness of the model was verified by comparing the values obtained with the theoretical and real values recorded on the tram, the latter obtained from a test campaign carried out on Hitachi vehicles in collaboration with an Italian transport company (ANM) that operates trams in the city of Neaples. The T-ADAS system is integrated into the vehicle logic by considering the driver's actions through the manipulator. When the designated braking action is not performed correctly, the system will adapt the braking curve to model one. In the final part, the implementation of the T-ADAS in the on-board network of the tram will be provided and the evaluation of the data traffic is performed. Infect, the newest railway network infrastructures based on Ethernet bus technologies can facilitate the integration of the T-ADAS systems with the Train Control Management System (TCMS), providing larger bandwidth and more flexible networks making this technology immediately transferable to railway systems such as the tram. Throughout the entire work of the thesis, the real braking data of the case study and implementation approach were provided by the ANM and Hitachi Rail S.p.A which the author is grateful for the contribution offered. [edited by Author]