Support Vector Regression Based Behavioral Model for Load Mismatched PAs

A new double-input double-output (DIDO) power amplifier (PA) behavioral model, which takes into account output load termination mismatches, is presented in this paper. The support vector regression (SVR) methodology, from the field of machine learning, is utilized to build this new model. The model is intended for the system-level simulation of amplifier chains subjected to wideband modulated signals. It is validated by comparing the predicted and measured data of a 15 W Gallium Nitride (GaN) based PA under a wide range of voltage standing wave ratios (VSWR). Compared with the conventional memory polynomial (MP) based DIDO models, the obtained results from the new model demonstrate the improved prediction capabilities.

[1]  C. Fager,et al.  Oscilloscope based two-port measurement system using error-corrected modulated signals , 2012, 2012 Workshop on Integrated Nonlinear Microwave and Millimetre-wave Circuits.

[2]  Jose C. Pedro,et al.  A New Double Input-Double Output Complex Envelope Amplifier Behavioral Model Taking Into Account Source and Load Mismatch Effects , 2015, IEEE Transactions on Microwave Theory and Techniques.

[3]  Corinna Cortes,et al.  Support-Vector Networks , 1995, Machine Learning.

[4]  Erik G. Larsson,et al.  Scaling Up MIMO: Opportunities and Challenges with Very Large Arrays , 2012, IEEE Signal Process. Mag..

[5]  Lingling Sun,et al.  Support Vector Regression-Based Behavioral Modeling Technique for RF Power Transistors , 2018, IEEE Microwave and Wireless Components Letters.

[6]  Christian Fager,et al.  Prediction of smart antenna transmitter characteristics using a new behavioral modeling approach , 2014, 2014 IEEE MTT-S International Microwave Symposium (IMS2014).

[7]  Ricardo Gonçalves,et al.  A new complex envelope behavioral model for load mismatched power amplifiers , 2017 .