Wind-Driven Rain Effects on Automotive Camera and LiDAR Performances

  • Author(s) / Creator(s)
  • Modern vehicles are equipped with Advanced Driver Assistance Systems (ADAS) that rely heavily on a variety of sensors such as LiDAR, RADAR, SONAR, and camera to capture surrounding traffic data. Sensor performance has been observed to degrade during adverse weather conditions as the signals are attenuated due to soiling on the external surfaces of the sensor. This poses huge risks to both occupants and pedestrians. In order to improve the safety of ADAS, it is essential to understand and quantify the effects of soiling on sensor signals. This paper investigates driving-in-rain scenarios, which are some of the most ubiquitous but hazardous weather conditions that cause accidents every year due to reduced driver and sensor vision. Rain was simulated in a controlled environment onto an automotive external camera and a single-point LiDAR in a wind tunnel. Perceived soiling characteristics when the vehicle is moving at different speeds under different natural rain intensities are considered and measured, including impact, intensity, velocity, and droplet size distribution. Sensor performance is evaluated based on image processing techniques and statistical analysis on signal data. The results show that camera image degrades with both increasing rain intensity and driving speed, whereas LiDAR signal only worsens during heavier rain conditions. However, driving faster when perceived rain intensity is kept constant improves the perception with faster signal recovery.

    Part of the Proceedings of the Canadian Society for Mechanical Engineering International Congress 2022.

  • Date created
  • Subjects / Keywords
  • Type of Item
    Article (Published)
  • DOI
  • License
    Attribution-NonCommercial 4.0 International