Using Graphical Processing Units for Enhancement of Metrology Systems

本研究的目的是了解如何处理高密度、高速传感器密集检测系统产生的大量数据。实现这一目标的方法是形成对计算计量学的新的基本理解，因为它涉及到加速计算机图形学中使用的计算几何技术。这个公式最终将在大多数计算机显卡中包含的图形处理器单元上实现。图形密集型计算机应用程序（如游戏）大大提高了这些显卡的性能，并使它们易于在大多数计算机系统上使用。新制定的计量算法将在这些系统上运行速度提高几个数量级。如果成功，这项研究的好处将包括提高检验和质量控制能力，为各种各样的制造商。受益最大的是那些产品精度高、产量大的制造商。特别是，以微型和纳米制造部门为目标的公司将能够有效和成本有效地监测和改进其工艺和产品。这种质量控制的提高最终将为美国制造商在包括MEMS、微纳米系统等新兴领域在内的各个领域带来更强大、更有竞争力的地位。

The objective of this research is to understand how to process large quantities of data being generated by high density, high speed sensor-intense inspection systems. The approach to accomplish this objective is to formulate a new fundamental understanding of computational metrology, as it relates to the computational geometry techniques used in accelerated computer graphics. This formulation will ultimately be realized on graphical processor units that are incorporated into most computer video cards. Graphically intensive computer applications such as games have driven a significant increase in the capabilities of these cards, and made them readily available on most computer systems. The newly reformulated metrology algorithms will operate orders of magnitude faster on these systems.If successful, the benefits of this research will include improved inspection and quality control capabilities for a wide variety of manufacturers. Manufacturers that will benefit the most are those whose product is high precision and high volume. In particular, firms targeting the micro and nano manufacturing sectors will be able to efficiently and cost effectively monitor and improve their processes and products. Such an increase in quality control will ultimately yield a stronger and more competitive position for manufacturers in the United States, in a variety of sectors including the emerging areas of MEMS, micro and nano-systems.