Ultra Precision Visual Servo Control of Micro Object

There is a constant demand for increasingly accurate and reliable precision motion control in advanced instrumentation and manufacturing processes requiring the highest precision. At precision levels of nanometer and beyond, precise positioning and/or alignment of micro objects in 3D space can be a daunting task prone to many sources of error. Better solutions to this problem can benefit areas such as photonic alignment, micro/nano assembly and nano instrumentation.;State-of-the-art motion stages can achieve motion resolution at sub-nanometer levels; however, due to non-linearity and/or sensor drift, maneuvering the stage with that level of accuracy throughout the entire workspace of the stage is extremely difficult if not impossible. A robust solution is to employ direct metrology which acquires real-time feedback from the object space of the application, enabling direct servo control which continuously compensates for time varying error sources. Visual feedback, employed to achieve direct metrology, can provide global information, which defines a common reference frame in the object space with respect to which multiple objects are registered, revealing the relative position and orientation among objects that are subject to positioning, alignment, and engagement.;The object of this thesis is to analyze the various aspects of and provide novel solutions to the task of achieving direct metrology and visual servo control applied to nanometer-scale motion control. A novel measurement technique is developed which enables six-degree-of-freedom motion measurement with nanometer precision. This measurement technology is integrated into a motion control scheme which performs positioning and alignment of micro objects throughout a 3D volume. Furthermore, a novel theoretical analysis of measurement bias in image-based motion estimation is developed which enables a bias compensation scheme. Finally, analysis of the sampling and exposure mechanism in the video capturing process to formulate a novel solution for the aliasing problem in visual motion sensors is presented.