2022 Volume 2022 Issue 1 Pages 120-128
Automated driving equivalent to Level 4 in urban areas requires advanced and autonomous recognition functions using on-board AI, as well as road facilities, communication facilities, and other infrastructure to support these functions. At the same time, the installation of road and communication facilities throughout Japan requires massive budgets, so it is necessary to study the minimum necessary infrastructure and recognition technologies. Therefore, in this project, for the purpose of advancing discussions of future cooperative areas, we conducted research and investigations on the recognition technologies and infrastructure that will be essential for automated driving systems through public road tests of automated vehicles, chiefly through universities and other organizations that are capable of publishing a certain amount of acquired data and the technologies used to acquire that data. Based on the knowledge gained from the public road tests, we investigated the factors and issues that cause recognition failures in (non-cooperative) automated driving systems, and evaluated the effectiveness of road traffic environment data provided by the infrastructure.