水下仿生移动双臂机器人的智能协调控制与规划方法

Underwater intelligent operation is one of the hot spots of underwater robot research, which is significant for development and utilization of ocean resources. This project aims at investigating the basic theory and key technologies of intelligent coordination operation of an underwater biomimetic mobile robot with dual arms for underwater tasks, such as collaborative grabbing, equipment docking. The main research topics are listed as follows. Firstly, considering coupling reduction and drag reduction, system model of the underwater mobile robot with dual arms will further be analyzed. The layout optimization method of four undulatory-fin units and dual arms will be given. Moreover, parameter mapping optimization model of multiple undulatory-fin units will be established. Then, anti-interference hovering control of the biomimetic robot based on deep neural network and extended state observer will be explored, and coordination control between biomimetic vehicle and dual arms based on feedforward compensation by estimation will be investigated. After this, coordinated operation planning of dual arms based on deep autoencoder will be explored to imitate the skills of operating persons. At last, numerical simulations and physical experiments will be conducted to verify the effectiveness and the practicability of the proposed methods by simulation system construction and robot platform reconstruction. The key technique will be accumulated by the implementation of the project. It is expected that the outcome of this proposal will provide important theoretical and technical support for intelligent coordination operation of underwater robot.

水下智能作业是水下机器人研究的热点问题之一，对海洋资源的开发与利用具有重要的意义。本项目针对水下协同抓取、设备对接等作业任务，拟以水下仿生移动双臂机器人为研究对象，深入探讨其智能协调作业的基础理论与关键技术。主要研究内容包括：综合考虑减耦、减阻等问题，深入分析水下仿生移动双臂机器人系统模型，给出四波动鳍、双作业臂机构的总体布局优化方法，建立多波动鳍参数映射优化模型；探讨基于深度神经网络方法和扩张状态观测器的仿生本体抗扰悬停控制，深入研究基于预估前馈补偿的本体-双作业臂协调控制方法；模拟操作人员遥操作技能，研究基于深度自动编码机的双臂机器人协调作业规划学习方法。最后构建仿真系统和改造机器人系统平台，对所提理论方法的有效性和实用性进行数值仿真分析和物理实验验证。通过项目的实施，积累关键技术，预期研究成果将为水下机器人智能协调作业提供重要的理论方法和技术支撑。

本项目研究了水下仿生移动作业机器人智能协调作业的基础理论与关键技术。结合计算流体力学技术和参数辨识方法，建立了水下仿生移动作业机器人的系统模型，采用微元分析法建立了波动鳍推力与波动鳍运动参数的映射模型；设计了用于目标识别与定位的图像处理模块、用于进行可控速度估计的跟踪微分器、建立图像特征误差及距离误差空间至期望可控速度空间映射关系的悬停控制器，实现了水下仿生移动作业机器人的自主悬停；给出了带有追踪微分器和状态观测器的任务优先运动学控制算法，利用自适应波动鳍参数调节方法实现了基于前馈补偿的仿生机器人运动控制，研究了仿生机器人-作业臂的协调控制方法；给出了一种基于任务参数隐半马尔科夫模型的共享控制方法，研究了双臂协调作业方法。在本项目资助下，相关理论与方法在IEEE Trans.等期刊发表或在线发表SCI论文10篇，申请5项国家发明专利。上述取得的研究成果和关键技术为水下机器人智能协调作业提供了必要的理论技术支持。