This work describes the development of RBOT, a robot-driven radio base station maintenance system. RBS deployment and maintenance tasks are increasing both in complexity and density with the introduction of 5G microcells. The main objective behind our application is to reduce maintenance costs by developing an integrated system based on a robotic arm to operate on the RBS front plane. This work details the challenges and solutions for maintaining the connectivity of networking cables in such an environment. More specifically, it discusses the problems of inserting communication cables and removing defective ones attached to an RBS equipment. We also examine the creation and evolution of a maintenance system from a simulation scenario to a real-world setup. RBOT’s interface allows both remote teleoperation and autonomous operation. It also contains Augmented Reality features providing a first-person view for remote teleoperation to increase environmental awareness. Furthermore, by applying convolutional neural networks for faulty cable classification, RBOT can actuate over the RBS and manipulate the cables with a unique, accurate, and robust gripper we specially designed for cable connector handling.

这项工作描述了机器人驱动的无线电基站维护系统 RBOT 的开发。随着 5G 微蜂窝的引入，RBS 部署和维护任务的复杂性和密度都在增加。我们的应用程序背后的主要目标是通过开发基于机械臂的集成系统来降低维护成本，以在 RBS 前平面上运行。这项工作详细介绍了在这种环境中保持网络电缆连接的挑战和解决方案。更具体地说，它讨论了插入通信电缆和移除连接到 RBS 设备的有缺陷的电缆的问题。我们还研究了维护系统从模拟场景到现实世界设置的创建和演变。RBOT 的接口允许远程遥控操作和自主操作。它还包含增强现实功能，为远程遥控操作提供第一人称视角，以提高环境意识。此外，通过将卷积神经网络应用于故障电缆分类，RBOT 可以在 RBS 上驱​​动并使用我们专为电缆连接器处理而设计的独特、准确和坚固的抓手来操纵电缆。