Ogden, UT · May 2, 2002 · Current Test Program Sets (TPS) are based on board models to duplicate the functionality of the Unit Under Test (UUT). Technical specifications from the original equipment manufacturer are used to generate pass/fail criteria, which are then compared to output measurements collected in a 'static' test sequence. TPS developed using board models have proven to be expensive to develop and maintain while being ineffective in many cases. The USAF needs to develop TPS with maximized effectiveness and minimal development and maintenance costs.
This research will investigate the feasibility of implementing a process to automate TPS development using COTS software to model/emulate UUT functionality. This research effort explores the integration of artificial intelligence technologies with software and hardware modeling/emulation to generate functional interface specifications for 'dynamic' TPS. The diagnostic effectiveness of the adaptive TPS will be demonstrated during the prototype phase.
The result of this research and development effort will significantly reduce the development and maintenance cost of Test Program Sets (TPS). The resulting methodologies and tools will introduce innovative ways to develop adaptive TPS. The dependability can be demonstrated directly and indirectly by reducing the cost of No-Fault-Found (NFF) cases due to the alignment of pass/fail criteria used in different levels of Unit testing. The software prototype will be modular allowing re-use by different TPS thus increasing software reliability and consistency. Testing procedures will be executed faster and more efficiently thus reducing the cost of maintenance and increasing system availability.
|
