Abstract: To address the coexistence of discrete topological evolution and continuous dynamic state variation during the on-orbit assembly of modular space structures (MSS), this paper investigates the dynamic modeling and integrated pose control problem for variable-topology systems. First, an assembly-process-oriented dynamic modeling strategy is proposed based on the deactivation-activation mechanism of structural modules, which enables a continuous description of the topology reconfiguration process. A unified reference frame is introduced to avoid frequent coordinate system reconstruction caused by center-of-mass variations. A coupled orbit-attitude-structure dynamic model is established using the first kind of Lagrange equations, with the Kelvin-Voigt contact model incorporated to characterize the impact dynamics during module docking. On this basis, to cope with system parameter jumps and multi-source uncertain disturbances during the assembly process, an integrated pose control strategy with control gains adaptively updated according to the assembly stages is proposed. An integral sliding mode controller compensated by a radial basis function (RBF) neural network is designed to achieve approximation and compensation of lumped disturbances. The stability of the closed-loop system is proven via Lyapunov theory. Numerical results demonstrate that the proposed control method can effectively handle the complex dynamic variations during the assembly process, and achieve high-precision stable control under various disturbance conditions. This study provides a theoretical basis for the dynamic modeling and control system design of the on-orbit assembly process of MSS.

Key words: on-orbit assembly, modular space structures, dynamic modeling, sliding mode control, neural network