音圈电机直驱的柔性解耦定位系统动力学特性与控制方法

For the development of micro-devices and their packaging manufacturing, the basic issues related to a voice coil motor direct-drive positioning system with flexible decoupling mechanism are studied. Considering the characteristics of the direct-drive of voice coil motor and flexible decoupling mechanism, a broadband dynamic modeling methodology in multi-energy domains for the high-speed precision positioning system with short stroke is proposed based on power bond graph and experimental identification method. The dynamic modeling and characteristics analysis are carried out, and the effects of the model parameters and disturbances on the dynamic behavior of the system are investigated. The autofocus micro-vision system is established using a voice coil motor. Through introducing adaptive genetic algorithm to image segmentation and geometric moments into visual localization, respectively, a micro-vision localization algorithm with high precision and high efficiency is presented based on image features. On the basis of the dynamic model, a discrete sliding mode controller with adaptive reaching law is developed. The high-speed precision positioning is realized through integration of the micro-vision system and motion system. The study can provide basic theory and methodology for the further development of such positioning system, meanwhile it has great significances for the improvement of the quality and efficiency of the micro-device packaging.

针对微器件及其封装制造的发展趋势，研究音圈电机直驱的柔性解耦定位系统相关基础问题。综合考虑音圈电机直驱形式以及柔性解耦的特点，提出一种基于功率键合图和实验辨识方法的短行程高速精密定位系统多能域宽频动力学建模方法，进行动力学建模与动态特性分析，考察模型参数以及干扰等对系统动态行为的影响规律。构建音圈电机驱动的具有自主调焦能力的显微视觉系统。将自适应遗传算法引入图像分割，将几何矩引入视觉匹配定位，探索一种基于图像特征的高效精密显微视觉匹配定位算法。以系统动力学模型为基础，设计一种基于自适应趋近律的离散滑模变结构控制系统。将显微视觉系统与运动系统集成，实现高速精密定位。该项研究为该类定位系统的进一步发展提供可供借鉴的理论和方法支持，同时对提高微器件封装制造的质量和效率具有重要意义。

结合微器件封装制造的发展趋势，针对显微视觉导引的音圈电机直驱柔性解耦定位系统相关科学和技术问题开展了研究工作。提出了基于机电联合仿真技术的精密定位系统设计方法，完成了定位系统的动态设计。综合考虑音圈电机直驱形式以及柔性解耦的特点，基于能量方法，利用功率键合图建立了短行程高速精密定位系统的动力学模型，以系统动力学模型为基础，设计了一种离散滑模变结构控制器，考察了控制效果。搭建了显微视觉系统，将自适应遗传算法引入图像分割，将几何矩引入视觉匹配定位，实现了一种基于图像特征的微器件精密显微视觉匹配定位算法。将显微视觉系统与运动系统集成，有效实现了显微视觉导引下的快速精确定位，系统定位误差在1.1微米之内，分辨率0.5微米。项目相关研究成果发表在IEEE，ASME等国内外高水平学术刊物上， 其中SCI检索期刊论文13篇，JCR 1区论文1篇，2区论文4篇。EI检索15篇，申报发明专利7项，培养研究生7名，积极开展国内外学术交流和合作。研究成果对提高微器件封装制造的质量和效率具有重要意义。