| Air-cushion vehicle(ACV)is a kind of surface ship with special performance.The ACV injects the gas into the bottom of the ship through the lifting system,resulting in a layer of air between the bottom and the water surface,and it does not have underwater devices,which is different from the ordinary surface ships.The special structure achieves the purpose of reducing the navigational resistance and increasing the speed of the ship.At the same time,weak anti-interference ability and poor stability problem was exposed in ACV.The interference of ocean breezes and sea waves may cause the ACV heeling and other dangerous situations,it may also cause the ACV overturning.So it is very necessary to study the control system of the ACV.In this paper,the heading-heeling decoupling control of ACV is mainly studied.The ship needs to change course according to the route and sea condition when the ACV sails in the water,so the rudder angle needs to be changed to control the heading.As the air rudder is usually located above the center of gravity of the hull,the rudder Angle changes often give a Rolling moment on the ACV.The wind and wave disturbances can also produce strong influence on the heeling of ACV.And the heeling will change the pressure of air cushion,and increase the instability of the ACV.Similarly,the work of the nozzle will give a Yawing moment on the ACV,this force also causes the change of course.All above shows that there is coupling between heading and heeling.In order to ensure the stable and safe navigation of the ACV,the air rudder and nozzle are required to work together.In order to facilitate the research of hovercraft,it is necessary to model the hovercraft mathematically.Firstly,the fixed and moving coordinates are established,and the conversion relationship between the two coordinates is solved.Secondly,it is necessary to analyze the force and establish the dynamic mathematical model of the ACV.The mathematical model of wave and wind is established to simulate the real ocean environment.The simulation experiment such as direct sailing,rudder rotation,and nozzle rotation is carried out,which demonstrates the sailing properties of ACV.In this paper,the PID neural network control method is applied to the heading-heeling decoupling control of ACV.The method is similar to multilayer feed-forward neural network,except that the hidden layer of PID neural network only has one layer,and whose transfer function is not a single function but consists of proportional,integral and derivative functions.This method not only has the characteristics of neural network self-learning but also has differential and integral dynamic characteristics.An improved method has been designed to deal with the shortcomings of neural networks,and a new method is proposed to select the initial value of connection weight.Then the PID neural network control method is used to control the heading of the ACV.A single-channel PID neural network control system is designed,and a classical PID control method is added as a contrasting item in the simulated experiment.Simulation results show that PID neural network control is better than classical PID control in ACV heading control.The next a multivariable PID neural network decoupling controller is designed,which is used in decoupling was used in decoupling control between heading and heeling.Heading control is a contrasting item in the simulated experiment.Simulation results show that the decoupling control method leads to the heading angle and the sideslip angle is smaller.It means that decoupling control makes the ACV have high stability and anti-interference. |