4DRaL: Bridging 4D Radar with LiDAR for Place Recognition using Knowledge Distillation

2026-03-27

AI Summary

This work addresses the challenge of robust place recognition under all-weather conditions using 4D millimeter-wave radar, which suffers from high noise levels and sparse point clouds. To overcome this limitation, the authors propose 4DRaL, a novel framework that, for the first time, transfers a high-performance LiDAR-to-LiDAR place recognition model to the 4D radar domain via knowledge distillation, enabling both radar-to-radar (R2R) and radar-to-LiDAR (R2L) cross-modal place recognition. The framework incorporates three key components—local image enhancement, feature distribution distillation, and response distillation—to effectively align cross-modal features and enhance student model performance. Experimental results demonstrate that 4DRaL achieves state-of-the-art accuracy in both R2R and R2L place recognition across normal and adverse weather conditions, significantly advancing the applicability of 4D radar for all-weather localization.

Abstract

Place recognition is crucial for loop closure detection and global localization in robotics. Although mainstream algorithms typically rely on cameras and LiDAR, these sensors are susceptible to adverse weather conditions. Fortunately, the recently developed 4D millimeter-wave radar (4D radar) offers a promising solution for all-weather place recognition. However, the inherent noise and sparsity in 4D radar data significantly limit its performance. Thus, in this paper, we propose a novel framework called 4DRaL that leverages knowledge distillation (KD) to enhance the place recognition performance of 4D radar. Its core is to adopt a high-performance LiDAR-to-LiDAR (L2L) place recognition model as a teacher to guide the training of a 4D radar-to-4D radar (R2R) place recognition model. 4DRaL comprises three key KD modules: a local image enhancement module to handle the sparsity of raw 4D radar points, a feature distribution distillation module that ensures the student model generates more discriminative features, and a response distillation module to maintain consistency in feature space between the teacher and student models. More importantly, 4DRaL can also be trained for 4D radar-to-LiDAR (R2L) place recognition through different module configurations. Experimental results prove that 4DRaL achieves state-of-the-art performance in both R2R and R2L tasks regardless of normal or adverse weather.

Problem

• 4D radar
• place recognition
• sensor robustness
• data sparsity
• adverse weather

Innovation

• 4D radar
• knowledge distillation
• place recognition
• LiDAR
• cross-modal learning