RLPR: Radar-to-LiDAR Place Recognition via Two-Stage Asymmetric Cross-Modal Alignment for Autonomous Driving
This work is significant for autonomous driving, as it improves all-weather localization by enabling radar-based localization using existing LiDAR maps, which is crucial for safety and reliability in diverse environmental conditions.
This paper addresses radar-to-LiDAR place recognition for autonomous driving, aiming to localize radar scans within existing LiDAR maps to enable all-weather autonomy. The authors propose RLPR, a framework that extracts structural features from radar and LiDAR data and employs a two-stage asymmetric cross-modal alignment strategy. RLPR achieves state-of-the-art recognition accuracy and strong zero-shot generalization across four datasets.
All-weather autonomy is critical for autonomous driving, which necessitates reliable localization across diverse scenarios. While LiDAR place recognition is widely deployed for this task, its performance degrades in adverse weather. Conversely, radar-based methods, though weather-resilient, are hindered by the general unavailability of radar maps. To bridge this gap, radar-to-LiDAR place recognition, which localizes radar scans within existing LiDAR maps, has garnered increasing interest. However, extracting discriminative and generalizable features shared between modalities remains challenging, compounded by the scarcity of large-scale paired training data and the signal heterogeneity across radar types. In this work, we propose RLPR, a robust radar-to-LiDAR place recognition framework compatible with single-chip, scanning, and 4D radars. We first design a dual-stream network to extract structural features that abstract away from sensor-specific signal properties (e.g., Doppler or RCS). Subsequently, motivated by our task-specific asymmetry observation between radar and LiDAR, we introduce a two-stage asymmetric cross-modal alignment (TACMA) strategy, which leverages the pre-trained radar branch as a discriminative anchor to guide the alignment process. Experiments on four datasets demonstrate that RLPR achieves state-of-the-art recognition accuracy with strong zero-shot generalization capabilities.