2019/12/03 by Eric Heiden, Ziang Liu, Heiden, Eric +5 · 1 citation
Physics and Astronomy · Engineering · Environmental Science · #Advanced Optical Sensing Technologies #Robotics and Sensor-Based Localization #Remote Sensing and LiDAR Applications
paper · pdf · doi:10.48550/arxiv.1912.01652
Light Detection and Ranging (LIDAR) sensors play an important role in the\nperception stack of autonomous robots, supplying mapping and localization\npipelines with depth measurements of the environment. While their accuracy\noutperforms other types of depth sensors, such as stereo or time-of-flight\ncameras, the accurate modeling of LIDAR sensors requires laborious manual\ncalibration that typically does not take into account the interaction of laser\nlight with different surface types, incidence angles and other phenomena that\nsignificantly influence measurements. In this work, we introduce a physically\nplausible model of a 2D continuous-wave LIDAR that accounts for the\nsurface-light interactions and simulates the measurement process in the Hokuyo\nURG-04LX LIDAR. Through automatic differentiation, we employ gradient-based\noptimization to estimate model parameters from real sensor measurements.\n