张帅军, 刘卫东, 李乐, 等. 基于RBF神经网络补偿的ROV运动控制算法[J]. 水下无人系统学报, 2024, 32(2): 311-319 doi: 10.11993/j.issn.2096-3920.2023-0033 引用本文: 张帅军, 刘卫东, 李乐, 等. 基于RBF神经网络补偿的ROV运动控制算法[J]. 水下无人系统学报, 2024, 32(2): 311-319 doi: 10.11993/j.issn.2096-3920.2023-0033 ZHANG Shuaijun, LIU Weidong, LI Le, LIU Jingbin, GUO Liwei, XU Jingming. ROV Motion Control Algorithm Based on RBF Neural Network Compensation[J]. Journal of Unmanned Undersea Systems, 2024, 32(2): 311-319. doi: 10.11993/j.issn.2096-3920.2023-0033 Citation: ZHANG Shuaijun, LIU Weidong, LI Le, LIU Jingbin, GUO Liwei, XU Jingming. ROV Motion Control Algorithm Based on RBF Neural Network Compensation[J]. Journal of Unmanned Undersea Systems , 2024, 32(2): 311-319. doi: 10.11993/j.issn.2096-3920.2023-0033 张帅军, 刘卫东, 李乐, 等. 基于RBF神经网络补偿的ROV运动控制算法[J]. 水下无人系统学报, 2024, 32(2): 311-319 doi: 10.11993/j.issn.2096-3920.2023-0033 引用本文: 张帅军, 刘卫东, 李乐, 等. 基于RBF神经网络补偿的ROV运动控制算法[J]. 水下无人系统学报, 2024, 32(2): 311-319 doi: 10.11993/j.issn.2096-3920.2023-0033 ZHANG Shuaijun, LIU Weidong, LI Le, LIU Jingbin, GUO Liwei, XU Jingming. ROV Motion Control Algorithm Based on RBF Neural Network Compensation[J]. Journal of Unmanned Undersea Systems, 2024, 32(2): 311-319. doi: 10.11993/j.issn.2096-3920.2023-0033 Citation: ZHANG Shuaijun, LIU Weidong, LI Le, LIU Jingbin, GUO Liwei, XU Jingming. ROV Motion Control Algorithm Based on RBF Neural Network Compensation[J]. Journal of Unmanned Undersea Systems , 2024, 32(2): 311-319. doi: 10.11993/j.issn.2096-3920.2023-0033 针对作业型遥控水下航行器(ROV)在模型参数不确定和外部环境干扰下的运动控制问题, 提出了一种基于径向基函数(RBF)神经网络的自适应双环滑模控制策略。首先, 对于ROV外环位置控制采用改进趋近律的积分滑模控制方法, 对于ROV内环速度控制采用指数趋近律的积分滑模控制方法; 其次, 为进一步改善滑模控制的抖振问题, 引入双曲正切函数作为滑模切换项; 然后, 利用RBF神经网络控制技术对ROV模型的不确定参数和外部扰动进行估计与补偿; 最后, 利用李雅普诺夫稳定性理论证明了整个闭环系统的稳定性, 并对作业型ROV的运动控制进行了数值仿真。仿真结果验证了所设计的控制器可以实现ROV航行的精确控制, 并能够有效抑制模型不确定参数和外部扰动对ROV运动的影响。

遥控水下航行器 /  运动控制 /  径向基函数 /  自适应双环滑模控制 / Abstract: In view of the motion control problem of the operation-type remotely operated vehicles(ROVs) under the uncertainty of model parameters and the disturbance of the external environment, an adaptive double-loop sliding mode control strategy based on radial basis function(RBF) neural network was proposed. Firstly, the integral sliding mode control method with an improved reaching law was adopted for controlling the position of the ROV’s outer loop, and the integral sliding mode control method with an exponential reaching law was adopted for controlling the speed of the ROV’s inner loop. Secondly, in order to further improve the chattering problem of sliding mode control, the hyperbolic tangent function was introduced as the sliding mode switching term. Subsequently, the RBF neural network control technology was used to estimate and compensate for the uncertain parameters and external disturbances of the ROV model. Finally, the stability of the whole closed-loop system was proved by using the Lyapunov stability theory, and the motion control of the operation-type ROV was simulated numerically. The simulation results verify that the controller designed in this paper can achieve precise control of ROV navigation and effectively suppress the influence of uncertain parameters of the model and external disturbances on ROV motion.

Key words: remotely operated vehicle /  motion control /  radial basis function /  adaptive double-loop sliding mode control /  neural network t′ 性能参数CSMCDSMCRBF-DISMC $ {t_x} $/s86.928.228.0$ {t_y} $/s67.651.147.9$ {t_z} $/s119.457.649.0$ {t_\psi } $/s47.836.029.6$ {{{M}}_{{\text{MAE}}}} $$ {x_m} $/m0.108 60.071 20.057 6$ {y_m} $/m0.075 50.122 60.054 1$ {z_m} $/m0.217 20.186 00.163 5$ {\psi _m} $/(°)$0.003\;2$$0.038\;0$$0.002\;7$$ {R_{{\text{RMSE}}}} $$ {x_r} $/m0.225 90.172 60.162 8$ {y_r} $/m0.133 80.170 30.115 4$ {z_r} $/m0.404 10.353 60.350 6$ {\psi _r} $/(°)$0.006\;7$$0.042\;3$$0.006\;6$