Autonomous driving systems are being vigorously developed worldwide and are expected to exhibit dramatic improvements in capability and widespread adoption in the future. To guarantee their safety and security, it is necessary to verify that sensors and systems function normally under diverse natural environments (rain, fog, light, snow, etc.). Anticipating the growing need for such verification, Mitsubishi Heavy Industries Machinery Systems, Ltd. has developed the indoor Adverse Weather Automated-driving Research Environment as an evaluation system for autonomous mobility. This report introduces a case where two rainfall environments with different raindrop size distributions despite having the same rainfall intensity were generated, and their impacts on LiDAR were compared; as a result, the reduction rate of reflection intensity varied depending on the rainfall environment, thereby demonstrating the importance of evaluation that considers detailed meteorological characteristic data.
