The process of identifying and handling critical scenarios of traffic situations in the OpenX format

Abstract

The process of identifying and processing critical scenarios (CS) of road situations plays an important role in the development of highly automated vehicles, as it improves their safety and adaptability to real-world operating conditions. Modern automatic control systems of highly automated vehicles require checking the correctness of the response to a wide range of traffic situations, including CS. Without a detailed analysis of the CS, it is impossible to ensure a high level of safety of the highly automated vehicles, especially in difficult conditions of urban and suburban traffic, in which there are completely different, but therefore no less dangerous situations. 

This paper discusses methods for detecting and processing traffic CS using the ASAM OpenX standard, including conditions, vehicle behavior, infrastructure parameters, etc. 

Special attention is paid to the use of traffic modeling tools such as Eclipse SUMO and esmini. SUMO is used to simulate traffic flows, analyze vehicle behavior, identify potentially dangerous maneuvers, and calculate key CS parameters. Based on the data obtained, scenarios were created in the OpenSCENARIO format, which were then reproduced in esmini to verify the behavior of vehicles in the CS. 

The proposed approach is aimed at supporting the process of substantiating the safety of highly automated vehicles within the framework of the SOTIF concept. It integrates into the architectural development stage of the system and is aimed at functional safety engineers who face difficulties in identifying and analyzing CS. Automated analysis of road scenarios using the proposed method helps to identify potentially dangerous situations, contributing to a more complete and reasoned justification of safety. This, in turn, forms the basis for improving the safety level not only of the automated system, but also of road traffic in general.