In the process of rectification of the waterway, it is necessary to lay a soft body row, which is a kind of engineering ship designed for the needs of the project. The operation mode of the paving ship is changeable, and the number of anchors used is random. The model-based method is difficult to model all working conditions, and the method based on the intelligent control algorithm has complex network structure and cannot adapt to complex working conditions. For the automatic control of paving ships, the core is to control the speed of each winch, and the collection of speeds of each winch constitutes a series of time series. In this paper, we use the advantages of LSTM neural network in processing time series data prediction, learn the advanced features of winch speed in each dimension, and establish a speed prediction model for deep learning. Therefore, the problem of complicated modeling and poor control effect caused by random use of anchors under complicated working conditions is solved.
