Research On Neural Network Control Method For Constant Gravity System Of Space Robot

At present,space exploration and space station construction in the aerospace field are becoming more and more important,and the development and control methods of space robots are also critical.The space robot constant gravity system studied in this paper is a system where the space robot simulates the movement of the manipulator under the low gravity or zero gravity environment in space on the ground.The system is also called a hanging type or hanging type control system,which is a three-dimensional control system composed of horizontal following and vertical constant tension.Horizontal follow-up control is controlled by two permanent magnet linear synchronous motors on the X and Y axes,that is,during the horizontal movement of the robot,the hanging trolley tracks the movement to achieve lowspeed and high-precision motion control;constant gravity control is controlled by the vertical direction.The electric cylinder and the winch motor are controlled together to achieve a zerogravity control mode that simulates space on the ground.First,through a large amount of literature research at home and abroad,an in-depth analysis of the current status of gravity unloading of the suspension system is carried out.Understand that the current constant gravity control method has the problems of complicated mechanical structure,small robot motion range,and insufficient consideration of control operation strategy.In view of these problems,the constant gravity control system studied in this paper,and the mathematical model of the constant gravity system is derived through the motor model and operating mode.Secondly,in order to solve the problems of poor robustness of the space robot's tracking motion in the horizontal direction,slow tracking response,and unstable follow-up control caused by the excessive horizontal swing angle of the hanging wire rope,this paper designed a fuzzy sliding mode controller for low speed and high speed.Precision tracking motion control.In order to improve the system robustness and reduce the response time,a sliding mode variable structure control method with good robustness,fast response speed and simple structure was selected.However,due to the large chattering of the sliding mode control method when the system is running,the accuracy cannot be achieved.Based on the above analysis,this paper will combine the fuzzy control algorithm to blur the slope of the sliding mode error,and simulate and analyze the fuzzy sliding mode control method by MATLAB/Simulink.Finally,for the problem of small moving range of space robots,this paper builds a composite control method of electric cylinder and winch motor to increase the control range;in order to meet the requirements of control accuracy,a three-input single-output reverse neural network control algorithm is designed;In order to improve the accuracy of the gravity unloading feedback results,the RBF observer is designed to provide real-time feedback on the pulling force and vertical position;In order to determine the relationship between the parameters of the control object and the effective control of the system,machine learning based on the RBF algorithm is used to perform network prediction analysis and determine the control law of the system.The mathematical model of loading tension/robot position conversion identified by orthogonal least squares is used to compare and analyze the input and feedback force/position.Finally,it is simulated by MATLAB,and the effectiveness of the control method is verified.