A Universal Machine-Learning-Based Automated Testing System for Consumer Electronic Products

A Novel System to Increase Yield of Manufacturing Test of an RF Transceiver through Application of Machine Learning

Electronic manufacturing and design companies maintain test sites for a range of products. These products are designed according to the end-user requirements. The end user requirement, then, determines which of the proof of design and manufacturing tests are needed. Test sites are designed to carry out two things, i.e., proof of design and manufacturing tests. The team responsible for designing test sites considers several parameters like deployment cost, test time, test coverage, etc. In this study, an automated test site using a supervised machine learning algorithm for testing an ultra-high frequency (UHF) transceiver is presented. The test site is designed in three steps. Firstly, an initial manual test site is designed. Secondly, the manual design is upgraded into a fully automated test site. And finally supervised machine learning is applied to the automated design to further enhance the capability. The manual test site setup is required to streamline the test sequence and validate the control and measurements taken from the test equipment and unit under test (UUT) performance. The manual test results showed a high test time, and some inconsistencies were observed when the test operator was required to change component values to tune the UUT. There was also a sudden increase in the UUT quantities and so, to cater for this, the test site is upgraded to an automated test site while the issue of inconsistencies is resolved through the application of machine learning. The automated test site significantly reduced test time per UUT. To support the test operator in selecting the correct component value the first time, a supervised machine learning algorithm is applied. The results show an overall improvement in terms of reduced test time, increased consistency, and improved quality through automation and machine learning.