Ranging and Obstruction Mapping using Ray Optics Approach to Model LIDAR Technology for Self-Driven Cars

Light Detection and Ranging (LIDAR) is a technology to range and image the surrounding obstructions. The range gives quantitative information to the Machine Learning (ML) algorithm about the obstacle’s distance from the LIDAR system. The image map gives the approximate height and width of the obstruction, assisting the ML algorithm in making a set of instructions to avoid the obstacle. In this paper, we discuss a geometric optics approach to model LIDAR system. The various control parameters of modeling LIDAR system such as cone release angle, the number of rays launched, and angular step size of elevated and azimuthal angle step are discussed. The effect of these control parameters on the error of mapped image are also discussed. The error of the mapped image of the car and pedestrian with 10,000 released rays, cone angle of 5°, and angular step size of 2° was reported with an error of 18% ± 3%.