Improved decoupling for maglev planar actuators with coil-to-coil misalignments: A LM iterative commutation tuning algorithm

The maglev planar actuator is a core component of ultra-precision motion stages in various industrial applications. Commutation serves as the critical first step in decoupling; however, its performance is usually limited by the mismatch between the actual and theoretical electromagnetic models. To address this issue, this paper proposes a data-based Levenberg-Marquardt (LM) iterative commutation tuning strategy. Unlike existing methods that consider only mover-to-stator misalignments, the proposed strategy models and tunes the phase errors of individual coils, thereby providing a more accurate commutation model. Moreover, the LM algorithm is employed to solve the resulting nonlinear least-squares problem, providing improved robustness and tuning performance compared with the gradient-descent (GD) algorithm. Experimental results on a moving-coil maglev planar actuator (MCMPA) show that the proposed strategy remains effective with tuning points selected along the X DOF, along the Y DOF, or distributed along both DOFs, and reduces the peak Z-DOF coupling error induced by X/Y-DOF movements by at least 47.7% under high-dynamic tracking. Furthermore, the proposed strategy outperforms the GD-based tuning method and the approach considering only mover-to-stator misalignments, while the tuned parameters can also be successfully applied to the direct decoupling model.