ReAL: Machine Learning Detection of Reflective Attacks Against Lidarometry

LiDAR-based technologies are rapidly becoming widespread, especially with the rise of autonomous vehicles (AVs) that rely on LiDAR to know what is around them at any given moment. This paper proposes a novel approach to mitigate reflective attacks on LiDAR systems by implementing a Support Vector Machine (SVM) model to detect specular surfaces. By incorporating machine learning with the sensor readings of the LiDAR, the risks associated with reflective surface interference are drastically reduced, ensuring continuous and reliable vehicle operation. This approach allows for LiDAR-based technologies to continue functioning properly despite obstacles that they cannot measure. The code is available on https://github.com/ChiefAj23/ReAL-ReflectiveAttack-Detection-Lidar.git.