Two degrees-of-freedom bellows-driven ultra-precision positioning platform using adaptive dynamic surface control

Abstract

This paper presents the design, analysis, and control of a compact ultra-precision positioning platform with two degrees of freedom, driven by bellows actuators and supported by aerostatic slideways. This configuration enables large-stroke motion under friction-free conditions, overcoming the conventional trade-off between travel range and positioning resolution. To address the output-feedback structure and inherent pneumatic nonlinearities, a radial basis function neural network is employed for efficient system approximation, and a neural network state observer is constructed to estimate unmeasurable system states. Dynamic surface control, as an advanced variant of backstepping, is further adopted to simplify the control design, while a tracking differentiator is incorporated to provide smooth signal filtering and avoid the complexity explosion typically encountered in derivative operations. The stability of the closed-loop system is rigorously established via a Lyapunov-based analysis. Extensive experimental results demonstrate that the proposed platform reliably achieves large-stroke ultra-precision positioning, validating the effectiveness of the control framework and confirming its potential for high-performance precision engineering applications.