【摘要】 针对水下机械臂在复杂水流环境下操作精度的提升问题，提出了一种基于自适应神经网络的滑膜控制策略。采用Newton-Euler法构建了适用于恒定均匀水流影响的双关节水下机械臂动力学模型，以提高模型精度并充分考虑实际应用中的环境因素。针对模型中存在的错误和不可预知的外部干扰，设计了融入自适应神经网络的滑膜控制器，有效补偿模型不准确带来的误差，抵抗多种类型的不确定外部扰动，从而显著改善了水下机械臂的轨迹跟踪速度和准确性。通过Matlab/Simulink平台进行仿真实验显示，相比于传统的PD滑模控制算法，提出的自适应神经网络滑膜控制方法在均匀水流干扰下，大大提高了水下机械臂的跟踪响应速度和精度，增强了系统在动态变化水流条件和模型不确定性面前的适应性。该策略的实施不仅提升了控制精度和响应速度，还保证了系统的稳定性，为水下机械臂在现代化海洋牧场等复杂水下作业环境中的应用提供了可行的解决方案。更多还原

【Abstract】 To address the problem of improving the operating accuracy of underwater robotic arms in complex water current environments, a slip membrane control strategy based on adaptive neural network is proposed. A kinetic model of a double-jointed underwater robotic arm applicable to the influence of constant and uniform water currents is constructed by using the Newton-Euler method in order to improve the model accuracy and fully consider the environmental factors in practical applications. In response to the errors and unpredictable external disturbances in the model, a sliding film controller incorporating an adaptive neural network is designed to effectively compensate for the errors brought about by model inaccuracies and resist multiple types of uncertain external disturbances, thus significantly improving the trajectory tracking speed and accuracy of the underwater robotic arm. Simulation experiments via the Matlab/Simulink platform show that, compared with the traditional PD sliding mode control algorithm, the proposed adaptive neural network sliding film control method greatly improves the tracking speed and accuracy of the underwater robotic arm under the uniform water current disturbances, and enhances the adaptability of the system in the face of dynamically changing water current conditions and model uncertainties. The implementation of this strategy not only improves the control accuracy and response speed, but also ensures the stability of the system, which provides a feasible solution for the application of underwater robotic arms in complex underwater operating environments, such as modernised sea ranches.更多还原