Machine Learning in a Quality Managed RF Measurement Workflow
Aric W. Sanders, John Bass, Arpita Bhutani, Mary A. Ho, James C. Booth
Advances in artificial intelligence, or more specifically machine learning, have made it possible for computers to recognize patterns as well or better than humans. The process of quality management in radio-frequency measurements is an arduous one that requires a skilled human to interpret complex data sets and frequently leads to long diagnostic routines when a measurement fails a quality check, or an operator makes a mistake. We demonstrate the application of machine learning classifiers to improve a common task in a quality managed measurement workflow. Specifically, we demonstrate that a machine learning classifier can predict the device attached to a measurement apparatus and additionally predict if one of the connections has been improperly torqued, furthermore predicting which connection is loosened.