Awareness of potentially dangerous situations for VRUs in a smart city

• Main Author: Teixeira, Pedro Veloso
• Publication Date: 2021
• Format: Master thesis
• Language: eng
• Source: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Download full: http://hdl.handle.net/10773/34010
• topic: 5G; Collision avoidance; Edge computing; Event-driven architecture; ITS; ITS safety; Smart city; Vehicular Ad-hoc networks; Vulnerable road users

According to Eurostat, in 2019, most – 70% - fatal road accidents in urban areas involved vulnerable road users (VRUs), such as children, impaired people, cyclists, and animals, with this percentage increasing in 2020 to at least 77% of fatal accidents. With the advent of Intelligent Transportation Systems (ITS), with vehicles being part of a Vehicular Ad-hoc Networks (VANET) and smart cities connecting all parts of an urban environment, new solutions for fighting against these accidents can be considered. Moreover, with the introduction of autonomous driving solutions, avoiding dangerous road situations and protecting the VRUs is more critical than ever. This dissertation proposes a multi-sensor solution for preventing potential accidents between vehicles and vulnerable road users. A smart city has several sensors, dispersed in the vehicles (and aggregated by the On-Board Units, OBUs), in the VRUs (e.g., smartphones and smartwatches), and in the road itself (e.g. cameras, radars). These different nodes communicate with each other through several wireless access technologies, most notable short range standards such as ITS-G5, C-V2X or long range technologies such as LTE, 5G and, in the future, 6G. By aggregating and processing such information, a system was implemented to predict and notify potential hazardous situations involving VRUs and vehicles. This system was based on the computation of risk zones and prediction of collision points between VRUs and vehicles. The current infrastructure from the Aveiro Smart City (thanks to the Aveiro Tech City Living Lab project) is leveraged to deploy and evaluate the system. Information from the city’s sensors and vehicles is gathered and joined to the information from the VRUs own devices in an hybrid architecture, with an edge and cloud deployment. The obtained results show that the system is promising at predicting potential collisions while pointing at some important deployment decisions that must be made to ensure proper notification timings, such as the usage of multi-homing, 5G, and 6G, and following the concept of edge computing. The developed system can be considered a step towards a level 5 of full automation of autonomous vehicles. The potential collision can be part of the pipeline for the decision of the following action.