Adaptive compensation control for ultra-large-inertia shaking tables based on motion-trend deviation

Ultra-large-inertia shaking table systems exhibit strong nonlinear disturbances, low damping, and extremely large equivalent mass, which significantly degrade control accuracy. To address these issues, this paper proposes an adaptive compensation control method based on generalized motion-trend deviation. First, a dual-channel state prediction method is developed by adaptively fusing a dynamics model channel for physical consistency with an online data-driven channel for localized nonlinearities via a dynamic confidence mechanism. Second, an online compensation controller using a Radial Basis Function neural network is designed to mitigate residual errors based on this deviation, which quantifies the dynamic mismatch between the inertial reference trajectory and the actual system response. Simulation and experimental results demonstrate that the proposed method effectively suppresses waveform distortion induced by friction and backlash. Compared with conventional controllers, the framework achieves a 33% reduction in maximum tracking error. The proposed strategy isolates nominal inertial dynamics from dissipative forces, providing a robust and proactive compensation structure for ultra-large shaking tables, although its application to multi-axis decoupling remains a subject for future research.