Hilti SLAM Challenge 2023: Benchmarking Single + Multi-session SLAM across Sensor Constellations in Construction

同时定位和建图系统是手持设备和机器人应用中定位的关键因素。过去几年，Hilti SLAM挑战赛成功地为一些全球最佳SLAM系统的高精度基准测试提供了平台。然而，这些系统的更多功能有待探索，例如跨不同传感器套件的平台不可知性和多会话SLAM。这些因素间接地成为在实际应用中具有强韧性和易部署性的指标。目前不存在一个公开可用的数据集和基准测试组合，同时考虑这些因素。Hilti SLAM Challenge 2023数据集和基准测试解决了这个问题。此外，我们提出了一种新颖的基准标记设计，用于从机器人上安装的现成LiDAR观测地面上的预测测点，并估计其毫米级精度的位置算法。挑战的结果显示，整体参与度增加，单次会话SLAM系统变得越来越准确，成功地在不同的传感器套件中运行，但相对较少的参与者执行多会话SLAM。

Hilti SLAM Challenge 2023 Dataset and Benchmark aims to explore the capabilities of simultaneous localization and mapping systems, including platform agnosticism and multi-session SLAM, which are important for real-world applications.

The challenge proposes a novel fiducial marker design and an algorithm for mm-level accuracy estimation of its position using an off-the-shelf LiDAR mounted on a robot. The results show an increase in overall participation, single-session SLAM systems getting increasingly accurate, successfully operating across varying sensor suites, but relatively few participants performing multi-session SLAM.

The Hilti SLAM Challenge 2023 Dataset and Benchmark provides a benchmark combination publicly available for evaluating SLAM systems with high accuracy and robustness. The proposed fiducial marker design and algorithm enable accurate positioning for real-world applications. The challenge results demonstrate the progress and potential of SLAM systems in various sensor suites and settings.

Recent related studies in the field of SLAM include 'A Review of Simultaneous Localization and Mapping' by C. Cadena et al. and 'Deep Learning for LiDAR-based Human Pose Estimation in Indoor Environments' by Y. Zhang et al.