Gaussian processes regression for optimizing harmonic impedance trajectories in continuous class-F power amplifier design

This paper proposes an advanced optimization-oriented strategy to design a continuous Class-F power amplifier (PA) using Gaussian processes regression (GPR). Optimal harmonic impedance trajectories for achieving high efficiency are obtained at a Cree CHG40010 package plane. GPR is exploited to optimize matching networks for fitting the desired trajectories of harmonic impedances. By combining GPR with Agilent's Advanced Design System (ADS), the optimization procedure runs automatically and predicts optimal design parameters which match well with the expected harmonic impedance trajectories. Measured results show that greater than 60% efficiency is attained over a 50% bandwidth from 1.5 to 2.5 GHz. The output power is larger than 40.2 dBm with gain varying from 13.2 to 15.4 dB across the band.