Dr-PoGO: Direct Radar Pose-Graph Optimization
Cedric Le Gentil, Weican Li, Leonardo Brizi, Timothy Barfoot
AI summary
Problem
Existing radar-based SLAM methods typically rely on feature or point cloud extraction, which discards valuable raw sensor information and struggles with robust loop-closure registration without a reliable initial guess.
Approach
Dr-PoGO leverages direct radar odometry for continuous tracking and introduces a coarse-to-fine loop-closure registration that combines feature-based RANSAC alignment with direct cross-correlation refinement, all integrated into a global pose-graph optimization.
Key results
- State-of-the-art trajectory accuracy over 300 km of real-world automotive data
- Robust loop-closure detection and registration in harsh weather and dynamic environments
- Novel coarse-to-fine direct registration pipeline eliminating the need for an initial pose guess
- Public release of implementation and dataset samples
Why it matters
Enables reliable all-weather autonomous navigation for ground vehicles by leveraging millimeter-wave radar's penetration capabilities with a highly accurate and robust SLAM pipeline.
Abstract
This paper introduces Dr-PoGO, a method for Simultaneous Localization And Mapping (SLAM) using a 2D spinning radar. Unlike cameras or lidars that require line-of-sight, millimetre-wave radars can ‘see’ through dust, falling snow, rain, etc. Accordingly, it is a great modality for robust perception regardless of the weather conditions. While most existing radar-based SLAM methods rely on the extraction of point clouds or features to perform ego-motion estimation, Dr-PoGO leverages direct registration techniques for odometry (DRO) and loop-closure registration. An off- the-shelf radar-focused place recognition algorithm, RaPlace, provides loop-closure candidates. As RaPlace does not provide relative transformations, Dr-PoGO introduces a coarse-to-fine registration that uses visual features and descriptors to obtain an initial guess for the direct transformation refinement. The global trajectory is optimized in a pose-graph optimization. Dr- PoGO demonstrates state-of-the-art performance over 300 km of data in various real-world automotive environments. Our implementation is publicly available: https://github.com/ utiasASRL/dr_pogo.