Inferential estimation ofhighfrequency LNA gainperformance using machine learning techniques

Functional testing ofradiofrequency integrated circuits isa challenging task andone that isbecoming an increasingly expensive aspect of circuit manufacture. Duetothedifficulties withbringing high frequency signals off-chip, currentautomated test equipment (ATE)technologies are approaching the limits oftheir operating capabilities as circuits are pushed tooperateathigher andhigher frequencies. This paper explores thepossibility ofextending the operating range ofexisting ATEsbyusingmachine learning techniques toinfer highfrequency circuit performance frommore accessible lowerfrequency andDC measurements. Results froma simulation study conducted on a lownoiseamplifier (LNA)circuit operating at2.4GHzdemonstrate thattheproposed approach hasthepotential tosubstantially increase theoperating bandwidth ofATE.