Localization of Autonomous Vehicle using 1D Automotive Radar Sensor

Car localization based on Radar and LIDAR Fusion

More Info
expand_more

Abstract

The overall purpose of this study is to establish novel methods of vehicle localization and mapping using a 1D linear automotive radar array in conjuncture with pre-existing lidar maps, and to test if the generated radar map can be made to be 3 dimensional. The reason for creating a novel localization method was to try and alleviate some of the weaknesses of existing localization techniques, such as GPS'(Global Positioning System) inability to localize in tunnels and other large structures, or lidar based positionings' low quality of service in certain weather conditions. The basic design of this study was to implement a SLAM(Simultaneous Localization And Mapping) system that co-registers radar data to radar data, and to then attempt to register radar data to lidar data, all taken from experiments with cars with these sensors driving around the TU Delft campus. After the execution of experiments, both on simulations and on real world data, it was established that it is in fact possible to localize the car by relating observed radar data to premade lidar maps, and to continually add to a cumulative map made with the radar data that can further aid the localization process. Furthermore, it can be observed that the radar map created using the 1D linear automotive array can be made to be 3D, though more experiments to establish the full potential of this capability are recommended.

Files