自主水下机器人系统感知–决策–控制技术综述

Review of Perception – Decision – Control Technologies for Autonomous Underwater Robotic Systems

摘要:

水下航行器–机械臂系统(Underwater vehicle-manipulator systems,UVMS)通过融合水下航行器的机动能力与机械臂的精细操作能力,为深海检测、维护与操作等复杂作业提供了技术支撑。本文围绕UVMS的关键技术体系,分析相关研究成果并开展综述。在感知方面,总结了水下环境感知方式、水下图像增强及目标检测技术,分析了其在复杂水下作业场景中的应用效果;在决策方面,综述了避障与路径规划以及运动轨迹规划方法,讨论了任务复杂性与环境不确定性对决策与规划性能的影响;在控制方面,梳理了UVMS的运动跟踪控制与力交互控制方法,比较了模型驱动、自适应及数据驱动控制策略在水下作业中的适用性。最后,从多模态鲁棒感知、智能决策与任务规划、运动–力协同控制角度探讨UVMS未来发展趋势并提出具有潜在应用前景的前沿技术。

Abstract:

Underwater vehicle-manipulator systems ( UVMS ) integrated the mobility of underwater vehicles with the dexterous manipulation capability of manipulators, thereby providing essential technological support for complex deep-sea operations such as inspection, maintenance, and intervention. A systematic review of recent research on the key technologies of UVMS was presented. Firstly, from the perspective of perception, underwater environment sensing methods, underwater image enhancement, and target detection techniques were summarized, and their application performance in complex underwater operational scenarios was analyzed. Secondly, from the decision-making perspective, obstacle avoidance, path planning, and motion trajectory planning methods were reviewed, with a discussion on the impact of task complexity and environmental uncertainty on planning and decision performance. Thirdly, from the control perspective, motion control and force interaction control methods for UVMS were examined, and the applicability of model-based, adaptive, and data-driven control strategies in underwater operations was comparatively analyzed. Finally, from the perspectives of multimodal robust perception, intelligent decision-making and task planning, and motion – force control, the future development trends of UVMS were discussed, and frontier technologies with potential application prospects were identified.