Active heave compensation of marine winch based on hybrid neural network prediction and sliding mode controller with a high-gain observer

The motion of the mother ship on the sea, especially the heave motion, greatly affects the normal operation of underwater equipment. In order to reduce the influence of mother ship heave motion on the normal operation of underwater equipment, a novel ship heave prediction algorithm and an active heave compensation (AHC) method based on sliding mode control (SMC) for electrically driven winch are suggested in this study. The principal contribution of this work is as follows: 1) An SMC with a high-gain observer is designed. Compared with the traditional SMC, the control accuracy is improved by 1.71%, 2.66%, and 2.16%, respectively, under Levels 2-4 sea states. To optimize the important SMC parameters, an improved particle swarm optimization strategy is applied. 2) In this study, a ship heave motion prediction method is proposed that fuses two neural networks of a convolutional neural network (CNN) and a long short-term memory network (LSTM). An improved grey wolf optimization approach is applied to optimize the CNN-LSTM's key parameters. The effectiveness of the proposed AHC control strategy is validated by conducting simulations and experiments using full-scale devices.