Course Keeping Controller Of Unmanned Surface Vessels Based On Quantum Neural Network

In this paper,the traditional BP neural network and RBF neural network are combined with the concept of quantum parallel computing to design two new quantum neural network models,which are quantum BP neural network model(QBP)and quantum RBF neural network model(QRBF).This kind of quantum neural network model combines the advantages of quantum and neural network.It has the advantages of both quantum parallel computing and neural network strong nonlinear ability.Among them,the ship autopilot is an important part of the research on the course keeping control of the unmanned surface vessels,and its performance is directly related to the navigation safety of the unmanned surface vessels.Therefore,it is very important to research and design a kind of quantum course keeping controller which can better guarantee the navigation safety of unmanned surface vessels.In this paper,this kind of quantum neural network model is applied to the research of course keeping control of unmanned surface vessels,and the course keeping control ability of quantum neural network model is proved by a series of simulation and physical simulation.The work of this paper is asfollows:In the first step,the classic BP neural network model and RBF neural network model are studied carefully,and their respective model structure,learning algorithm and operation mechanism mode are studied deeply.In this paper,the structures of two kinds of neural networks are introduced,the convergence performance and operation mechanism of two kinds of neural networks are analyzed theoretically.In the second step,we introduce the basic knowledge of quantum computation,and introduce the concept of quantum register to transform the input value of real state into quantum state which can be represented by quantum register through neural network.Then the neural network Transform the input value to an n-bit quantum superposition state,At the same time,for the hidden layer of neural network,we can also use the concept of quantum register to express,so we can adjust it by unitary operator,in this way,the parallel operation characteristic of neural network is used to improve the convergence speed of neural network.In the third step,a simple and easy to operate nonlinear PID controller is designed as the teacher controller of the quantum neural network controller.In this paper,a simple sine function combined with traditional PID is used to build a nonlinear PID controller for course keeping.In the fourth step,through a series of simulation experiments and real ship experiments,the results show that QRBF is better than QBP in convergence speed and robust performance for the course keeping control ability of unmanned surface vessels.