【作者】 金翠云；

【导师】 靳世久；

【作者基本信息】 天津大学 ， 精密仪器及机械， 2004， 博士

【摘要】 微机电系统(MEMS)是指集微型机械、微型执行器以及信号处理和控制电路、接口、通信和电源等于一体的可批量制作的微型器件或系统。由于其体积小、可批量生产、集成度高等优点,近几年得到了快速的发展并已步入产业化阶段。在MEMS的产业化进程中,测试技术越来越得到国内外许多MEMS研究机构的重视,其中MEMS动态特性测试技术在MEMS研发和产业化过程中具有最为重要的意义。本文基于机器微视觉技术,对MEMS动态特性中的平面微运动特性测试技术开展了研究,研究内容主要包括:1、系统地调研和分析了微机电系统、微机电系统测试技术及微机电系统平面动态特性测试技术的研究现状、应用领域和发展趋势。2、创新性地构建了可以同时满足两种测量原理:模糊图像合成原理和频闪动态成像原理的MEMS动态测试系统。从而可以根据测试要求和目的进行灵活选择。基于虚拟仪器LabVIEW开发软件测试平台,同时集成了MATLAB、VC等应用软件,扩充系统功能。3、提出了一种联合运用空域增强、小波域增强和空间灰度矩亚像素定位三种技术的MEMS运动轨迹提取算法,实现了MEMS器件运动轨迹的精确提取。在连续光照明条件下,基于模糊图像合成原理应用该算法对MEMS平面运动特性进行分析。通过扫频测量和扫压测量得到MEMS器件的平面运动特性。4、提出了基于块匹配与相位相关的亚像素精度的MEMS二维运动估计新算法,实现了时域测量和频域测量相结合的MEMS平面运动快速、精确估计。基于频闪成像原理,运用自行设计的频闪同步控制系统捕捉MEMS器件瞬间清晰的运动图像序列。利用该算法对MEMS器件的运动历程做分析,得到特定驱动频率下MEMS器件的幅度-相位曲线,还分析了MEMS器件的相频特性。利用扫频测量原理,获得了MEMS器件的幅频特性。5、提出了基于标号场和邻域优化法相结合的光流算法分析MEMS器件的运动历程,实现了MEMS平面内平动与旋转的精确测量。通过测量,得到了MEMS器件在特定驱动频率下的幅相特性;同时结合扫频测量,获得了MEMS器件的幅频特性。对两种基于频闪成像原理的二维运动估计方法做了比较和分析,对于不同的测试器件,根据测试要求可以做不同的选择。6、根据动态特性测试结果,分析和测量MEMS器件的力学特性参数。通过MEMS器件谐振频率推导MEMS微梁的杨氏模量和残余应力等力学特性。更多还原

【Abstract】 Micro-Electro Mechanical System (MEMS) is a kind of micro-device thatcombines micro-mechanism, micro-actuator, circuit for signal process and control,interface, communication and power supply. It contains many benefits such as smallvolume, batch process and highly integrated level. It develops very soon and stepsinto industry in recent year. In the course of developing, testing technique becomesmore and more important and many MEMS research departments attach to it. MEMSdynamic testing technique is the most important in all of MEMS testing requirements.This dissertation researches on in-plane micro-motion measurement for MEMS basedon machine micro-vision. The research work includes six aspects as follows.1. The application field, current status and development trends of Micro-electromechanical system, testing technique of MEMS and in-plane dynamic testingtechnique of MEMS are fully investigated and analyzed.2. MEMS dynamic testing system, which combined blur image synthetictechnique and stroboscopic dynamic image technique, is set up to measure in-planemotion of MEMS. Software testing plane is established based on virtual instrumentLabVIEW and also integrated with MATLAB and VC to extend system function.3. MEMS motion tracing extracting algorithm combined with airspaceenhancement, wavelet space enhancement and space gray level matrix sub-pixellocation technique is brought forward. Extracting MEMS motion trace accurately isrealized through it. In the condition of continue lighting, MEMS in-plane motioncharacteristics are analyzed using the algorithm based on blur image syntheticprinciple. The in-plane dynamic parameters are obtained through sweep frequencymeasurement and sweep voltage measurement.4. MEMS two-dimension motion estimation algorithm based on block matching,phase correlation and sub-pixel location technique is brought forward. EstimatingMEMS in-plane motion quickly and accurately is realized combined time space andfrequency space measurement. Applying stroboscopic synchronous control systemthat is self-made, MEMS clear motion image sequences are acquired. MEMS motionprocess is analyzed using the algorithm. The amplitude-phase curve of MEMS inspecial driving frequency is got. The phase-frequency character is also analyzed. Andthe amplitude-frequency curve is acquired through sweep frequency measurement.5. Optical flow algorithm combination of labeling field and neighborhoodoptimization is proposed to realize precise measurement for MEMS translation androtation. The amplitude-phase curve of MEMS in special driving frequency is got.And the phase-frequency character also analyzed associated with sweep frequencymeasurement. The two motion estimation methods based on stroboscopic imagingprinciple are compared. According to testing requirement it could choose any of themin practical application.6. The mechanical properties of MEMS are analyzed according to the testingresults of dynamic characteristics. The young’s modulus and residual stress ofmicro-beam are deduced by natural frequency of MEMS.更多还原