压电式微夹钳的智能建模和控制方法研究

In order to improve the intelligent level and control effect of the piezoelectric microgripper, this project focuses on the intelligent modeling and control for microgripper to achieve adaptation to varying tasks along with the dynamic coupling hysteresis nonlinearity. The research consists of the following contents. First, to satisfy the requirements of large stroke, high precision and high bandwidth, the topology design and multi-objective parameter optimization are adopted to design the compliant mechanisms. Then, a voltage-displacement-force coupling hysteresis nonlinear model and its online evolution method are researched based on the feedforward multilayer neural network, considering the influence of input frequency, temperature and load. On the basis of accurate modeling, the nonlinear intelligent control algorithm is studied. Based on the iterative learning identification method, the direct inverse feedforward control, which can be evolved on-line without the need of off-line identification, is designed to compensate the dynamic coupling hysteresis nonlinearity. The finite-band impedance feedback control is designed, and the method of on-line control parameters adjustment is studied. The full-band design limitation of control algorithms is overcomed. At last, the piezoelectric microgripper experimental platform is developed to verify the effectiveness of the method with various working conditions. This project aims to break through the technical bottleneck in piezoelectric microgripper, and provide a complete set of intelligent modeling and control methods for piezoelectric microgrippers, achieving its micro/nano precision control with various working conditions.

本项目以通过智能方法来提升压电式微夹钳的控制效果和智能化程度为目标，围绕动态耦合磁滞非线性和对不确定的操作任务的适应性展开压电式微夹钳的智能建模和智能控制研究，具体研究内容：综合考虑大行程、高精度和高带宽的要求，利用拓扑结构设计法和多目标优化法设计柔性机构；基于前向多层神经网络，结合具体工况，研究可精确描述微夹钳随输入特性变化、温度变化和负载变化的电压-位移-力耦合磁滞非线性模型及在线进化方法；在精确建模的基础上，研究非线性智能控制算法。基于迭代学习辨识法，设计无需离线辨识、可在线进化的直接逆前馈控制，补偿动态耦合磁滞非线性。设计有限频带阻抗反馈控制，研究控制参数在线调整方法，突破控制算法基于全频带设计的局限性；开发压电式微夹钳实验平台，结合多种工况，验证方法的有效性。项目旨在突破压电式微夹钳中存在的技术瓶颈，最终提供一套完整的智能建模和控制方法，实现其多工况下的微纳米级精密驱动控制。

随着物质在微观尺度特征研究的深入以及微机电系统的快速发展，研究和开发能替代操作人员双手、具有灵活操作微小物体能力的微夹钳，已成为越来越被关注的问题。本项目以通过智能方法来提升压电式微夹钳的控制效果和智能化程度为目标，围绕机构设计、磁滞非线性建模和鲁棒控制展开研究。综合考虑大行程、高精度和高带宽的要求，研究了微夹钳机构设计和几何参数优化，实现了45.58的超高放大比和683.7μm的超大行程。研究了率相关磁滞非线性智能建模与模型在线进化方法，解决了传统磁滞模型求逆复杂、计算/实验负担重、且率无关的难题，实现了率相关磁滞逆补偿器的直接求解，为微夹钳自适应鲁棒控制提供了信息基础。研究了非线性自适应鲁棒控制与参数智能调谐算法，解决了传统微夹钳控制基于全频带设计的难题，提升了微夹钳的控制性能及对不确定工况的适应能力。研制出压电式微夹钳实验样机，结合多种工况进行了验证。