Underwater docking technology enables autonomous underwater vehicles (AUVs) to execute long-term observation missions by periodically recovering and recharging AUVs. The conventional AUV homing and docking operations utilize acoustic and optical sensors at different ranges relative to the docking station. However, this method cannot perform perfectly in confined water regions because of the acoustic reflection and multipath effect. Thus, this paper proposes a novel navigation system, which fuses downward-looking visual odometry and model-based velocity for homing, and recognizes and tracks the light marker for terminal docking, in order to overcome the defects of the conventional navigation method. The reservoir experiment result verifies the effectiveness of the proposed method and shows good potential to extended applications in underwater routine cruise.

水下对接技术使定期航行的水下机器人（AUV）定期回收和充电，从而可以执行长期的观察任务。常规的AUV归位和对接操作使用相对于对接站处于不同范围的声音和光学传感器。然而，由于声反射和多径效应，该方法不能在受限水域中完美地执行。因此，本文提出了一种新颖的导航系统，该系统融合了向下看的视觉里程表和基于模型的速度进行归位，并识别和跟踪用于终端对接的光标记，以克服传统导航方法的缺陷。