Adaptive Neural Network Control for Spacecraft Attitude Takeover Based on Distributed Observers

For the attitude tracking control problem of multiple service satellites taking over non-cooperative spacecraft，with factors such as unknown model parameters，actuator faults and external disturbances considered，a neural network adaptive control method based on distributed observers is proposed for spacecraft attitude takeover. The nonlinear dynamic model with unknown parameters is approximated by the radial basis function （RBF）neural network. A distributed state observer based on neural network observers is employed to address the challenge where only partial satellites have access to target measurements，achieving observation consistency for the combined spacecraft under unknown model conditions. An adaptive compensation control law is designed to adaptively adjust control parameters according to actuator faults， enabling tracking control of reference attitude motion. The proposed method is applied to the attitude takeover problem of non-cooperative spacecraft，and simulation results demonstrate that precise attitude tracking control of the combined spacecraft can be achieved.