Abstract: Aiming at large and medium spatial experimental objects with complex motion, an engineering system composed of machine, motor and air-bearing is designed to simulate the spatial microgravity environment. Horizontal movement adopts air suspending, vertical movement adopts motor and reducer, constant force closed loop system is adopted by control system to track target gravity. For test object, friction non-linearity of machine system, outside disturbance and low speed of aircraft are considered. Neural sliding model equivalent controller with learning ability is adopted for adaptive learning and for compensation for various uncertain influences. Test result shows that the designed test equipment not only features light-weight structure and easy use, but also can compensate non-linearity dynamic influences of manipulators and achieve high-precision simulation of microgravity environment. So it has important reference value for three-dimensional spatial microgravity simulation test.

Key words: Ground test system, Neural network control, Sliding model control, Spatial microgravity