Research On Control Algorithm Of Three-floating Gyroscope Based On Neural Network

The Three-Float Gyroscope is very important in the fields of military strategic instruments,aerospace equipment and so on.For the development of Three-Float Gyroscope,China started nearly 30 years later,so far,the accuracy of China’s Three-Float Gyroscope is 2-3 orders ofmagnitude behind that of foreign countries.Therefore,improving the accuracy and stability of the Three-Float Gyroscope is still the top priority in the industry.In this thesis,the magnetic levitation control system of Three-Float Gyroscope is further studied in order to design a more accurate active magnetic levitation control algorithm.Based on the existing technical level of Xi ’an Aerospace Institute 16 and by reading a large number of scientific research materials,the control system of Three-Float Gyroscope is determined as the research content.In view of the existence of multiple input and output variables in the dynamic operation of the Three-Float Gyroscope.In addition,the coupling relationship between variables is uncertain,and the system model is complex and cannot be accurately determined.The influence of different control algorithms on the magnetic levitation system is analyzed.The work of this thesis is as follows:(1)The working principle of Three-Float Gyroscope and the output torque mechanism of magnetic levitation system are analyzed.And this system is described as a whole.Then the position detection,central control,output torque and float components of the system are analyzed theoretically and the model is built.(2)The motion law of gyro float under the interference of external force is analyzed.Three-dimensional float simulation model is controlled by PID.Then analyze and study different decoupling control algorithms suitable for multivariable control systems,and compare them through simulation experiments.A high precision magnetic levitation control algorithm is selected,which can not only solve the influence of coupling relationship between various directional variables on the system accuracy,but also realize the adjustable control parameters.(3)In order to make thealgorithm more fit with the actual demand,this thesis proposes a MPIDNN control algorithm based on the control environment.Through training,the model is self-adjusted to obtain the optimal solution of the control parameters.And the MPIDNN’s high efficiency is proved.Finally,the experiment was carried out on Three-Float Gyroscope test table to carry out the experiment,The drift precision of the fixed position of gyro is improved.The seesaw effect has lessened.The feasibility of the algorithm is fully verified.Through simulation and data comparison,in this thesis in the Three-Float Gyroscope control system,it fully verifies the excellent performance of the neural network PID control algorithm proposed.Multivariable coupling control of the control system is solved.It provides a good idea for further improving the precision of Three-Float Gyroscope.