基于LSTM神经网络的半潜式机器人垂直面控制研究

Research on Vertical Plane Motion Control of Semi-Submersible Vehicles Based on LSTM Recurrent Neural Network

摘要:

在简要分析半潜式机器人作业特点的基础上，指出了半潜式机器人垂直面运动控制的难点，介绍了近水面航行器相关控制技术。系统地论述了半潜式机器人基本控制原理以及反馈控制方法，提出了一种基于长短期记忆循环神经网络的半潜式机器人垂直面运动控制策略，给出了该控制策略的构建方法，并通过试验验证了该控制策略的有效性。

Abstract:

In this paper，the working characteristics of semi-submersible vehicles are analyzed briefly. The difficulties of vertical plane motion control are pointed out，and the control technologies related to near-surface vehicles are introduced. Then，the basic control principles and feedback control methods of semi-submersible vehicles are explained systematically. After that，the vertical plane motion control strategy of semi submersible vehicles based on long short-term memory recurrent neural networks is put forward. Finally，the construction method of the control strategy is provided，and the effectiveness of the control strategy is verified through tests.