Resarch On Orbit Service And Control Technology Of Space Manipulation

With the progress of space technology,the level of research,development and application of space is constantly improving.Countries around the world have launched a large number of spacecraft to meet the needs of various space missions.The structure of spacecraft is becoming more and more complex,and the development level is becoming more and more mature.The development of space control technology with rendezvous and docking,on orbit maintenance services,spacecraft orbit change,space debris avoidance as the main contents can greatly extend the service life of spacecraft,reduce the cost of spacecraft development and launch,and promote the development of human space industry.With the increase of space control tasks,space robots have been successfully applied to the on orbit service tasks of space shuttle,space station and satellite,assisting or even replacing astronauts in space environment,completing space control tasks,improving the safety of on orbit operation and service,and reducing risks.Its functions mainly include spacecraft life extension,space garbage cleaning,on orbit assembly of super large spacecraft and so on.In order to enhance the ability of space robot to deal with various uncertain conditions and tasks,it is of great significance to study the on orbit service control technology.Firstly,this paper analyzes the development status of space robot at home and abroad,the control theory research status of space robot and the research status of deep reinforcement learning technology.Then,the mathematical model of space robot is established;Euler angle is used to describe the attitude of satellite,and the mathematical model of attitude dynamics of rigid satellite is established;The kinematics and dynamics equations based on Lagrange equation of space robot under unconstrained and attitude controlled conditions are derived respectively,which provides a theoretical basis for the next simulation research.Secondly,the design and analysis of the satellite arm cooperative controller are carried out;The simulation platform of the space robot is built,the coordinate system of the space robot system is bound,and the pose control algorithm of the end of the space robot is studied.However,under unconstrained conditions,the satellite base of space robot will drift seriously,which is not conducive to the development of subsequent on orbit service tasks.Therefore,the satellite arm cooperative controller is designed,the simulation research is carried out,and its shortcomings are analyzed,and the solutions are proposed.Thirdly,the attitude stabilization controller based on deep reinforcement learning is designed;Due to the coupling effect of the space manipulator on the satellite base,the traditional PID attitude stabilization algorithm can not complete the compensation of the disturbance when it is disturbed.An attitude stabilization controller based on model free deep reinforcement learning is designed,and its internal mechanism and network structure are analyzed.The trained control strategy is used in simulation to verify the feasibility of the method.Finally,the end force controller of space robot is designed;The position based impedance control is studied,and the simulation test is carried out to verify the effectiveness of the classical impedance control algorithm,and its shortcomings are found.Aiming at the shortcomings of classical impedance control algorithm,particle swarm optimization algorithm is used to find the optimal impedance parameters,and the steady-state error of impedance control is analyzed and compensated.According to the dynamic characteristics of typical on orbit service tasks,the on orbit assembly mobile task and on orbit screw rotation task are simulated and analyzed.