Active Multi-Sensor Fusion for Inspection of Microelectronics

The fabrication of microelectronic components is a critical capability for both the national defense and economic security of the United States. Inspection processes are crucial in the fabrication of these components, assuring the economical production of high quality, state-of-the-art components. Current inspection processes for subminiature electronic assemblies typically rely on a single sensor and a single line of processing to detect defects in components. However, single-sensor processes are limited by equipment sensitivity and reliability. The fusion of evidence from multiple sensors can enhance the performance of sensor-based processes. For example, research and development in military situation assessment has produced a significant number of general research results and application frameworks. The objective of the proposed project is to design and document an implementable, functional fusion architecture for inspection of microelectronic components to increase the probability of detecting defects, reduce the incidence of false alarms, and allow discrimination between various types of defects. This active, multi-sensor fusion architecture will incorporate a real-time, on-line cycle of information gathering and evidence combination, as well as the off-line capability to assess historical data and form models for combining evidence.