Abstract

Robust and reliable speed sign detection is essential for realizing safe automated driving functions. Incorrect or non-recognition of speed signs can lead to excessive speed and, thus, danger in road traffic or even accidents if an automated system adapts the longitudinal guidance of the vehicle. This work conducted performance tests on speed sign recognition and adaptive speed adjustments of automated driving functions. For this purpose, a real-road test route was defined in the Ingolstadt area, which contains speed signs that could be critical for speed sign recognition. The tests were conducted with three vehicles from the manufacturers Tesla, Mercedes, and Volkswagen. Three tests were driven with each car during the day and at night to investigate daytime performance differences. The test route also allows the identification of possible infrastructural weaknesses in the installation of speed signs. The result shows that the systems correctly detected the applicable speed in more than 90% of the cases. Also, no significant differences in performance were found between day and night. However, the adaptive speed adjustment still needs to be improved since the speed was correctly adjusted at the beginning of the zone in only one-third of the cases.