Recent advancements in construction robotics have significantly transformed the construction industry by delivering safer and more efficient solutions for handling complex and hazardous tasks. Despite these innovations, ensuring safe robotic navigation in intricate indoor construction environments, such as attics, remains a significant challenge. This study introduces a robust 3‐dimensional (3D) robotic mapping and navigation method specifically tailored for these environments. Utilizing light detection and ranging, simultaneous localization and mapping, and neural networks, this method generates precise 3D maps. It also combines grid‐based pathfinding with deep reinforcement learning to enhance navigation and obstacle avoidance in dynamic and complex construction settings. An evaluation conducted in a simulated attic environment—characterized by various truss structures and continuously changing obstacles—affirms the method's efficacy. Compared to established benchmarks, this method not only achieves over 95% mapping accuracy but also improves navigation accuracy by 10% and boosts both efficiency and safety margins by over 30%.

中文翻译：

---

复杂施工环境中的高效 3D 机器人映射和导航方法


建筑机器人技术的最新进展为处理复杂和危险的任务提供了更安全、更高效的解决方案，从而显著改变了建筑行业。尽管有这些创新，但确保在复杂的室内建筑环境（如阁楼）中安全的机器人导航仍然是一项重大挑战。本研究介绍了一种专为这些环境量身定制的强大 3 维 （3D） 机器人映射和导航方法。该方法利用光检测和测距、同步定位和映射以及神经网络，生成精确的 3D 地图。它还将基于网格的寻路与深度强化学习相结合，以增强动态和复杂施工环境中的导航和避障能力。在模拟阁楼环境中进行的评估（以各种桁架结构和不断变化的障碍物为特征）肯定了该方法的有效性。与已建立的基准相比，该方法不仅实现了超过 95% 的测绘准确率，而且导航精度提高了 10%，效率和安全裕度提高了 30% 以上。