Research Of The High Precision Control System For Semi-active Microgravity Ground Simulation Platform

China's space industry has entered a new period, from the traditional single satellite launch to do communications, information collection into high challenging tasks for spacecrafts such as capture, docking, operation, maintenance. If a simulation microgravity environment on the earth is more similar to the space environment, the dynamic characteristics of the spacecrafts can be tested in the simulation environment before they are launched. And then the reliability of the spacecrafts can be made sure.Firstly the research background and the significance of microgravity simulation were described, then the microgravity simulation methods at home and abroad were anlyzed and compared, and according to the research outputs of our laboratory previously, a new semi-active space simulated microgravity system was constructed. Air bearing and hydrodynamic ball bearing were used to realize 5-DOF floating object. To realize 6-DOF floating object simulation, a vertical air cylinder was used to compensate its gravity. In vertical direction, active air floatation system couldn't satisfy the precision and dynamic performance of the system because of the nonlinear friction and air fluctuating, so the motor ball screw structure was designed to overcome the friction and residual gravity of the object.The force sensor was used to measure the value of the external force that the objects suffered. According to the value of the external force, the acceleration of the simulated object was estimated. The expected velocity at next time was evaluated by the least square method. The object was moving at the vertical direction by the ball screw structure, ensuring that the output of the force sensor equals to the gravity of the object, at the end, the dynamical behavior of the object has been simulated if it is in microgravity environment. Define the difference between the output of the force sensor and the gravity of the object as the uncompensated friction at the vertical direction, which value must be lower than one newton. After considering the interrupt and analyzing the error of the system, the controlling algorithm was improved. The improved controlling algorithm based on the discrete Kalman filter that fused two informations. The two informations are the suffered force and the velocity of motor seperately measured by the force sensor and the photoelectric encoder. The real experiment results show that the new controller is better than orignal one, the precision and the dynamic performance of the system were improved. The motors all have deadzone, and in the experiments the motor must change the direction frequently, so the deadzone has a large effect on the system. Because the ball screw has nonliner friction, the width of deadzone is asymmetric. The paper described all kinds of deadzone compensation methods, and then anlyzed the control characteristic of the motor, finally a fuzzy logic compensator was designed. By simulation the deadzone of the system was compensated, the dynamic performance has furtherly improved.