Currently, many levels of autonomous driving are being rapidly developed and are near deployment. No human intervention is required above level 4 of autonomous driving. Imperatively, a method for preventing and reproducing accidents that may occur during autonomous driving is desirable. In this study, the state change that occurs in V2X (vehicle to everything) environment is treated as first-class data to examine the cause by reproducing the accident when an accident occurs. We propose a framework that reproduces and restores state using event sourcing techniques that store first-class data. CCS Concepts • Computer systems organization ➝ Architectures ➝ Other architectures➝ Data flow architectures.
The vehicle-to-everything (V2X) environment comprises a variety of devices that exchange or share vehicle-to-vehicle, vehicle-to-infrastructure, and vehicle-to-pedestrian information while the vehicles are being driven. In V2X environments, large amounts of unstructured data are recorded by various devices. Storing large amounts of unstructured data in a relational database would require all the data generated by the different types of devices to be normalized. In addition, the transaction-based Atomic, Consistency, Isolation, and Durability (ACID) characteristics of relational databases would not be suitable for processing large amounts of data of this nature. In this study, we apply the Command and Query Responsibility Segregation (CQRS) pattern to separate commands and queries and to process large-scale data efficiently. NoSQL, which provides specialized functions for handling unstructured data, as well as the commands and queries, can be separated; thus, NoSQL and RDBMS can be selectively used depending on the V2X environment and system characteristics. We propose a data processing system suitable for the V2X environment using a CQRS pattern and the NoSQL-based database. By applying the CQRS pattern, NoSQL and RDBMS can be used together. When storing large amounts of data, use NoSQL, and when providing information or statistics to users, select an RDBMS to use together. In addition, both repositories can be easily scaled up when processing large amounts of data, making efficient use of resources.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.