Contact Accident Prevention System around Snowplows utilizing LiDAR and Machine Learning Technologies

Abstract

In heavy snowfall areas, the snow removal work by snowplows plays a significant role in securing transportation for local residents. However, there are many pedestrians and cars approaching the snowplows unintentionally, and the snowplow operators should pay attention to avoid the contact accidents with them, which reduces the efficiency of the snow removal work. As a result, it takes a long time to clean the road to keep the safe and secure life of local residents. Therefore, in this study, we develop a new contact accident prevention system around snowplows that detects pedestrians and cars approaching snowplows and notifies the snowplow operator in real-time. In this system, the sensor node installed on the snowplow analyzes the 3D point cloud data obtained by the LiDAR (Light Detection and Ranging) to detect the existence of the pedestrians/cars around the snowplow and notifies the snowplow operator of the results in real-time. In addition, the proposed system adopts the system structure of edge computing so that the system can be used in environments where high-speed mobile communications are not available (e.g., mountainous areas) and the system cannot leverage computing resources on the Internet. Furthermore, a machine learning method is utilized for quickly detecting pedestrians and cars with high accuracy from the 3D point cloud data obtained by the LiDAR.