Research On Adaptive Heading Control Of Wave Glider

Wave glider is a new type of ocean observation platform that can convert wave energy into forward kinetic energy,which has the characteristics of long range,long duration,no emission and no pollution.Excellent heading control capability is the core of the navigation control for ocean observation platforms like the wave glider.Nevertheless,the power of the wave glider comes from wave energy,which makes its own speed unstable.the ocean environment is complex and changeable,which makes improving the heading control ability of the wave glider is increasingly a hot spot and difficult point of related research.To address the above problems,this paper establishes a wave glider wave-driven speed prediction model,and combines BP neural network with PID controller and ADRC controller respectively to build an adaptive heading controller applicable to wave gliders for the characteristics of wave glider speed changes with wave factors.The main research contents are as follows.Firstly,this paper establishes the kinetic model and kinematic model of the "Black Pearl" wave glider;in order to reduce the influence of noise on the wave glider's heading control,the particle filtering process is performed on the heading control input of the wave glider,in view of the noise interference in the environment and in the heading sensor itself.Secondly,since the wave glider ground speed can be regarded as the combined speed of current speed and wave-driven speed,the sea trial experiment is designed to exclude the effect of current speed,and the mapping relationship between wave-driven speed and wave factor of wave glider is studied.Due to the nonlinearity of the mapping relationship,the prediction model of wave-driven speed is established using support vector regression to provide a basis for optimal wave glider heading control.Thirdly,for the changes of the environment in the wave glider and the perturbation of its own structural parameters,the BP neural network is introduced on the basis of the traditional PID algorithm,and a control method combining neural network and the traditional PID is used to dynamically adjust the PID parameters by using the nonlinear expression and self-learning ability of the neural network,and by adjusting the output range of the neural network,it is adapted and applied to the wave glider.For the problem that the BP neural network is easy to fall into the local minimum,the adaptive differential evolution algorithm is used to find the initial weights of the BP neural network,and the effectiveness of the algorithm is verified through simulation and experimental to prove that the algorithm can effectively improve the heading control ability.Fourthly,to address the defects of the PID such as error value jump caused by expectation value change,lack of differential extraction means,and lack of disturbance observation means,we introduce the active disturbance rejection control instead of the traditional PID control,and construct the expanded state observer for the perturbation of model parameters caused by the wave glider speed change.At the same time,the BP neural network is used to adjust the state error feedback law parameters online,which further improves the control quality and control accuracy,and is demonstrated by simulation and experiment.