Neural Network Motion Tracking Control of Piezo-Actuated Flexure-Based Mechanisms for Micro-/Nanomanipulation

This paper presents a neural network motion tracking control methodology for piezo-actuated flexure-based micro-/nanomanipulation mechanisms. In particular, the radial basis function neural networks are adopted for function approximations. The control objective is to track desired motion trajectories in the presence of unknown system parameters, nonlinearities including the hysteresis effect, and external disturbances. In this study, a lumped-parameter dynamic model that combines the piezoelectric actuator and the micro-/nanomechanism is established for the formulation of the proposed approach. The stability of the control methodology is analyzed, and the convergence of the position-and velocity-tracking errors to zero is proven theoretically. A precise tracking performance in following a desired motion trajectory is demonstrated in the experimental study. An important advantage of this control approach is that no prior knowledge is required for not only the system parameters, but also for the thresholds and weights of the neural networks in the physical realization of the control system. This control methodology is very suitable for the implementation of high-performance flexure-based micro-/nanomanipulation control applications.