Abstract: Aiming at the requirements of ground zero gravity simulation for spacecraft on orbit servicing missions, the zero gravity motion simulation technology based on industrial robot is studied. A force perception prediction model based on BP neural network is established. The model uses the attitude, acceleration, angular velocity and angular acceleration of the robot end as the input layer parameters, and the six dimensional force sensor data of the robot end as the output layer parameters. The high precision dynamic force perception of the robot end load is realized. The orthogonal experiment method is designed to determine the motion path points of the robot for sample data acquisition, and the force perception prediction model is realized to cover the whole workspace of the robot. Furthermore, based on the force perception data of the end load of the robot, the dynamics theory is applied to calculate the motion speed of the load in the state of weightlessness, and the robot is controlled to perform the corresponding motion, and the zero gravity motion simulation of the end load of the robot is realized.

Key words: On orbit servicing, Zero gravity simulation, BP neural network, Robot control