Abstract
The biaxial gantry system is extensively used in modern industrial applications that demand high productivity, machining precision, and smooth operation. Meeting these stringent requirements necessitates time-optimal, high-accuracy contour tracking with minimal vibration. However, parametric uncertainties, uncertain nonlinearities, and system constraints fundamentally challenge the concurrent realization of these objectives. This article proposes a novel two-layer control framework that seamlessly integrates online trajectory planning with adaptive robust control (ARC). The upper layer employs a nonlinear-filter-based planner to generate time-optimal trajectories along a given contour in real time, while explicitly enforcing jerk and input constraints to ensure smooth and feasible motion. The lower layer precisely defines contour error using the global task coordinate frame and employs ARC to accurately estimate and compensate for model uncertainties and suppress uncertain nonlinearities, thereby ensuring precise contour tracking. A key innovation lies in the closed-loop integration between layers: parameters identified online by the ARC are fed back to refine the planning model, enhancing trajectory planning results. Theoretical analysis demonstrates that contour tracking errors remain bounded, parameters converge to true values under persistent excitation conditions, and all system constraints are satisfied. Experimental results on an industrial biaxial gantry platform validate the proposed method, demonstrating superior contour accuracy, faster transient response, and reduced vibration compared to existing approaches.
Funding Information
  • Zhejiang Provincial Natural Science Foundation of China (LR23E050001)
  • National Natural Science Foundation of China (52575076)
  • The Seed Fund Cultivation Project of Ocean College, Zhejiang University (2025LJ002)

This publication has 35 references indexed in Scilit: