面向月球基地设备交互场景的目标检测与路径规划算法研究

随着国际月球科研站计划的快速推进，针对月球基地相关技术的研究受到了广泛关注。面向月球基地场景下不同的设备间进行自主接触式交互这一应用背景展开研究。考虑不同的设备在进行自主接触式交互时，需要准确地完成相互识别，并且规划出畅行无阻且较优的行进路线以保证往返活动顺利进行，基于Webots搭建了月球基地仿真环境进行目标检测和路径规划相关实验。以前后配备高精度RGB相机的交互设备为载体实现基于TPH-YOLOv5月球基地场景下的目标检测。此外，设计并实现了交互设备在月球基地场景下基于双向快速扩展随机树（Bidirectional RTT）的路径规划算法，以保证交互设备在月球基地环境下的相关工作顺利进行。仿真结果显示，该目标检测及路径规划算法能够实现预期功能。

Research on target detection and path planning algorithms for lunar base exploration scenarios

With the rapid advancement of the international lunar scientific research station program, research on lunar base-related technologies has received widespread attention. This article focuses on the application background of autonomous contact interaction between different devices in the lunar base scenario. Considering that different devices need to accurately identify each other when performing autonomous contact interaction and plan a smooth and optimal route to ensure smooth round-trip activities, a lunar base simulation environment based on Webots is built to conduct target detection and path planning experiments. The interactive device equipped with high-precision RGB cameras in the front and rear is used as the carrier to perform target detection in the lunar base scenario based on TPH-YOLOv5. In addition, a path planning algorithm based on Bidirectional RRT for interactive devices in the lunar base scenario is designed and implemented to ensure the smooth progress of related work of interactive devices in the lunar base environment. Simulation results show that the target detection and path planning algorithms used can achieve the expected functionality.