Compact neural network-enhanced synthesis of a novel decoupled 2-DOF tilt stage based on graded auxetic and contracting honeycomb

Abstract

Given the increasing demands for large strokes with low coupling for micro/nanomanipulation, a novel compact tilt stage is proposed based on the graded auxetic metamaterial to provide two degrees-of-freedom (DOFs) rotation motions in this paper. Firstly, graded auxetic and contracting structured honeycombs are designed with parallel combinations to provide amplified displacements driven by piezoelectric stacks, applying the orthogonal layering configuration with spatial distribution to alleviate rotation coupling. Compared with other tilt stages, larger rotation stroke and lower rotation coupling are yielded with compact dimension, attributed to intensive honeycomb structures. Then, the attention mechanism is involved in a compact neural network model to predict boundary conditions of the elliptical integral model, improving accuracy and simplifying derivation. Through sensitivity analysis and parametric optimization, the structural parameters are determined. To validate the effectiveness of design and modeling, the prototype of the proposed 2-DOF tilt stage is fabricated with the volume of 29*29*32 mm3. Experimental results show the rotation strokes of 17.98 and 18.82 mrad, with the coupling ratio of 0.46% and 0.45%, respectively. Graphical abstract: [Figure not available: see fulltext.]