Automated micro assembly has been identied to be a critical technology in the two emerging elds of micro systems technology and nano technology. No commercial success of either of these technologies is possible without reliable and cost-eective assembly and packaging technologies for the nished products at micro and nano scales.
Meanwhile, the availability of high resolution cameras and powerful microprocessors have made it possible for the vision systems to play a key role in solving the above problem. Vision techniques have been successfully implemented in the micro domain for 2D and 3D localization, object orientation determination and force estimation.
In this report, various state-ofart vision based assembly systems are briey described and an analysis of the same is presented. The vision algorithms in place in the current systems, can be broadly classied as segmentation based, color based, edge based, neural network based, optic ow based, model based and probability based (condensation, kalman lters) iterative techniques.
The report concludes with the formulation of the problem aimed at vision guided automated micro assembly system, to be tackled during the course of this project and plans towards achieving the same.
Micro assembly is the assembly of objects with micro scale and/or meso scale features under micro scale tolerances [YN04a]. The assembly and packaging of integrated circuits and the associated microelectronic systems have reached a stage of maturity of being close to full automation [PNH95]. These systems include inkjet printer heads and readwrite heads for information storage systems and the inertia sensors for automotive airbag deployment systems.
It can be noted that the basic requirement of an automated micro assembly is that the system must be able to transport parts and components of micro scale and be able to manipulate them so that precise spatial relation (with micro scale tolerances) can be established. For example, in die alignment and parts insertion or in packaging processes such as die bonding, device sealing, etc., the operating tolerances are very small and therefore, assembly and positioning tasks require very high precision and repeatability.Also, the assembly techniques for the macro domain do not directly translate to the micro domain because of the dominance of eects like Van der Waals forces, surfacetension, and electrostatic forces.
These forces are dicult to model accurately and are highly dependent on the operating environment. Thermal growth can introduce significant changes in dimensions of micro structures and cannot be compensated for unless the environment is regulated precisely. This indicates a clear need for sensor-based assembly of hybrid micro devices.
Vision systems compensate the lack of force sensors [Fur99] [PHL00] but are really bulky causing strong constraints on the design of micro manipulation systems.
Microscope optics and imaging unit
Project Done By Kaushik J