LiDAR-Based Place Recognition For Autonomous Driving: A Survey

Yongjun Zhang,Pengcheng Shi,Jiayuan Li

2023-07-29

Abstract:LiDAR-based place recognition (LPR) plays a pivotal role in autonomous driving, which assists Simultaneous Localization and Mapping (SLAM) systems in reducing accumulated errors and achieving reliable localization. However, existing reviews predominantly concentrate on visual place recognition (VPR) methods. Despite the recent remarkable progress in LPR, to the best of our knowledge, there is no dedicated systematic review in this area. This paper bridges the gap by providing a comprehensive review of place recognition methods employing LiDAR sensors, thus facilitating and encouraging further research. We commence by delving into the problem formulation of place recognition, exploring existing challenges, and describing relations to previous surveys. Subsequently, we conduct an in-depth review of related research, which offers detailed classifications, strengths and weaknesses, and architectures. Finally, we summarize existing datasets, commonly used evaluation metrics, and comprehensive evaluation results from various methods on public datasets. This paper can serve as a valuable tutorial for newcomers entering the field of place recognition and for researchers interested in long-term robot localization. We pledge to maintain an up-to-date project on our website <a class="link-external link-https" href="https://github.com/ShiPC-AI/LPR-Survey" rel="external noopener nofollow">this https URL</a>.

Robotics

What problem does this paper attempt to address?

The paper attempts to address the application of LiDAR-based Place Recognition (LPR) in autonomous driving. Specifically, the paper aims to:

1. **Fill the Research Gap**: Despite the extensive research on vision-based place recognition (VPR) methods in recent years, there is a relative lack of systematic reviews specifically targeting LiDAR-based place recognition (LPR). This paper aims to fill this gap by providing a comprehensive review.

2. **Overview of Existing Methods**: The paper conducts an in-depth review of existing LiDAR-based place recognition methods, including traditional handcrafted feature methods and advanced deep learning techniques. The authors categorize these methods into 7 classes and describe their advantages and limitations in detail.

3. **Provide Detailed Classification**: The paper offers a detailed classification and introduction of methods based on local descriptors and global descriptors, including handcrafted methods and learning-based methods. Additionally, it provides rich charts and tables to help readers understand open-source codes, backbone networks, feature types, similarity measures, memory costs, training strategies, operational efficiency, and localization capabilities.

4. **Summarize Datasets and Evaluation Metrics**: The paper summarizes existing datasets and commonly used evaluation metrics, and conducts a comprehensive comparison of various methods on multiple public datasets.

5. **Future Directions**: The paper proposes potential future research directions, providing valuable references for new researchers entering the field and scholars interested in long-term robot localization.

Overall, this paper aims to provide a comprehensive and systematic review of LiDAR-based place recognition research, promoting further research and development in this field.