Harnessing the Power of Ray Tracing for Enhanced 3D Map-based Localization using NanoVDB

In the evolving realm of autonomous mobile robots, accurate self-localization remains essential for successful applications in logistics, inspection, and maintenance in varied environments. While these robots optimize operations and ensure human safety, their effective functioning demands high precision in spatial localization, which can be challenged by external disruptions. To counteract this, 3D map-based localization and scan matching with LiDAR captured point clouds, has emerged as a vital tool. Leveraging the novel application of ray tracing, traditionally used in graphics rendering, this study innovates in the domain of 3D map-based localization. Given the computational intensity of using whole environment maps for point cloud registration, extracting localized point clouds become crucial. Addressing this, we introduced a ray tracing module, exploiting the advantages of the NanoVDB library. This research evaluates this module's efficiency for 3D map-based localization. The results highlight the module's proficiency in enhancing the Spot robot's localization precision to 0.148 meters, while also offering superior computational efficiency relative to conventional approaches.