RF-Power Amplifier Modeling and Predistortion Based on a Modular Approach

This paper presents the black-box modeling of a microwave power amplifier (PA) operating near saturation and the performance of a modular predistortion model designed for this PA. A 16-QAM modulation scheme has been used as input signal. Nonlinearities caused by the amplifier operating at low input back-off (IBO), that is close to compression, can change the output constellation and lead to difficulties in finding the correct model parameters since the black-box identification procedure is based on measured input/output signals. A segment of the input/output measurement data is processed to generate an initial model. Then pseudo-inverse techniques are used to find a parsimonious Wiener model that is cross-validated with the entire measurement data. The presented modeling approach results in a model intended to be numerically robust and having a high identification percentage based on the normalized mean squared error (NMSE) figure of merit. Once the PA model is obtained, a Hammerstein based predistorter (HPD) is estimated by means of the indirect learning approach (PA input-output measured data). Simulation and experimental results are provided in order to show linearity improvement achieved by this HPD. In band and out of band distortion compensation are measured in terms of adjacent channel power ratio (ACPR) and error vector magnitude (EVM) reduction