Performance Evaluation of 3D Mapping and Navigation using LiDAR and Automotive Radars in Inland Waterways Environment

Kurzfassung

This presentation investigates how automotive radars can be used for 3D mapping in inland waterways (IW), a setting where radar-based mapping has so far received limited attention. It introduces a method for building 3D maps with a network of four automotive radars installed on a water surface vessel (WSV), using data collected across various IW scenarios in Germany. The resulting radar-based maps are assessed against LiDAR-based maps via Iterative Closest Point (ICP), examining metrics such as fitness score and inlier correspondence RMSE. Findings show that radar-derived maps successfully represent the global geometry of waterways and surrounding infrastructure, though with less fine detail than LiDAR. Nonetheless, radar’s resilience to challenging environmental conditions, its suitability for detecting dynamic objects, and its cost advantages underscore its promise for improving autonomous navigation and transport operations in IW settings.