Anti-Wind and Wave Interference Control of Sea Gangway Based on DDPG Algorithm

Aiming at the control problem of ship-borne sea gangway under the in interference of wind and waves, the kinematic model of gangway is built by D-H parameter method, and the kinematic equation of the gangway is deduced. Then, the reward function sets the target point guidance, self-collision detection and training round limit as rewards and punishments, introduces the deep reinforcement learning DDPG algorithm to learn the action-reward model, and uses the learning results to manipulate the movement of the gangway in pursuit of higher reward value. Finally, a model is built in the simulation platform gazebo, and wind and wave interference is introduced into the Marine System Simulator (MSS) toolbox to train the model. The results show that after a certain number of training episodes, the end of the gangway can be stabilized in the specified position. Experimental results show that the convergence speed and efficiency of this method are good.