A recursive neural networks model for ship maneuverability prediction

In this paper, a recursive neural networks model is developed and applied to simulate the maneuvers of a tanker, which is full and may have inherent poor coursekeeping ability. In the present model, component force modules is developed to calculate five component forces as inputs to the networks. It consists of the net thrust and lateral force due to propeller revolution and rudder angle, approximate Munk moment, longitudinal component and lateral component of centrifugal force acting on ship hull. These forces are related to the input control variables such as ruder angle, propeller revolution and the output state variables such as motion velocities by very simplified functions without any undetermined hydrodynamic coefficients or empirical factors. The present recursive neural network is constructed with one input layer, one output layer and two hidden layers. Not only the above-stated forces, but also the outputs of longitudinal velocity, lateral velocity and yaw rate are fed back to the input layer of the network. In this study, an existing ship maneuvering simulation program, which has been developed basing on Japan MMG hydrodynamic model, is used for generating all the sample data of maneuvers for training and validating the recursive neural networks. The ship maneuvering motions are investigate d using the recursive neural networks which has been trained on limited maneuvers including turning, zigzag, spiral as well as accelerating maneuvers, and its validity to predict the maneuverability of a full ship with poor course stability is also discussed