Kim Jisoo, Park Bum-Jin, Roh Chang-Gyun, Kim Youngmin
Department of Highway & Transportation Research, Korea Institute of Civil Engineering and Building Technology, Goyang 10223, Gyeonggi-do, Korea.
Sensors (Basel). 2021 Nov 10;21(22):7461. doi: 10.3390/s21227461.
The performance of LiDAR sensors deteriorates under adverse weather conditions such as rainfall. However, few studies have empirically analyzed this phenomenon. Hence, we investigated differences in sensor data due to environmental changes (distance from objects (road signs), object material, vehicle (sensor) speed, and amount of rainfall) during LiDAR sensing of road facilities. The indicators used to verify the performance of LiDAR were numbers of point cloud (NPC) and intensity. Differences in the indicators were tested through a two-way ANOVA. First, both NPC and intensity increased with decreasing distance. Second, despite some exceptions, changes in speed did not affect the indicators. Third, the values of NPC do not differ depending on the materials and the intensity of each material followed the order aluminum > steel > plastic > wood, although exceptions were found. Fourth, with an increase in rainfall, both indicators decreased for all materials; specifically, under rainfall of 40 mm/h or more, a substantial reduction was observed. These results demonstrate that LiDAR must overcome the challenges posed by inclement weather to be applicable in the production of road facilities that improve the effectiveness of autonomous driving sensors.
激光雷达传感器在降雨等恶劣天气条件下性能会变差。然而，很少有研究对这一现象进行实证分析。因此，我们在对道路设施进行激光雷达传感时，研究了环境变化（与物体（路标）的距离、物体材料、车辆（传感器）速度和降雨量）导致的传感器数据差异。用于验证激光雷达性能的指标是点云数量（NPC）和强度。通过双向方差分析测试了指标的差异。首先，NPC和强度都随着距离的减小而增加。其次，尽管有一些例外情况，但速度变化并未影响这些指标。第三，NPC的值不因材料而异，每种材料的强度遵循铝>钢>塑料>木材的顺序，不过也发现了例外情况。第四，随着降雨量的增加，所有材料的两个指标均下降；具体而言，在降雨量达到40毫米/小时或更高时，观察到大幅下降。这些结果表明，激光雷达必须克服恶劣天气带来的挑战，才能应用于提高自动驾驶传感器有效性的道路设施生产中。
