Power Amplifier Modeling Comparison for Highly and Sparse Nonlinear Behavior Based on Regression Tree, Random Forest, and CNN for Wideband Systems

Mon, 01 Jan 2024 00:00:00 +0000

A Comparison of Surrogate Behavioral Models for Power Amplifier Linearization under High Sparse Data

Sat, 01 Oct 2022 00:00:00 +0000

Spectral Optimization and DPD for 5G Massive MIMO Systems

Sat, 01 Oct 2022 00:00:00 +0000

Overview

During my postdoctoral tenure at Tecnológico de Monterrey (2018–2026), supported by a CONACyT research fellowship, I led the development of DPD algorithms and spectral modeling techniques for 5G massive MIMO transmitter arrays, where per-antenna PA nonlinearity and crosstalk effects degrade system-level spectral efficiency.

Key Contributions

Developed surrogate behavioral models for PA linearization under high sparse data conditions, comparing polynomial, spline, and machine learning approaches (regression trees, random forests, CNNs) for wideband nonlinear modeling.
Designed DPD algorithms optimized for massive MIMO architectures, addressing the challenge of scaling linearization across large antenna arrays while managing computational complexity.
Performed system-level spectral efficiency analysis quantifying the impact of PA nonlinearity on massive MIMO capacity, beam pattern distortion, and adjacent channel interference.
Published results in Sensors (2022), IEEE ISCAS (2022), and contributed a book chapter to Machine Learning for Complex and Unmanned Systems (CRC Press, Taylor & Francis, 2024).
Advised 3 M.Sc. students and served on Ph.D. thesis committees related to this research line.

Tools & Technologies

MATLAB, Python (scikit-learn, TensorFlow), Keysight ADS, RF measurement testbed, statistical modeling frameworks.

Impact

This work bridges classical RF linearization theory with modern machine learning, demonstrating that data-driven surrogate models can match or exceed conventional polynomial DPD approaches — particularly in scenarios with limited or irregularly sampled training data, which is common in multi-band and multi-antenna deployments.

Reliable comparison for power amplifiers nonlinear behavioral modeling based on regression trees and random forest

Sun, 01 May 2022 00:00:00 +0000

SVM Classifier and evaluation of muscle power of EMG signals and Python implementation

Sun, 01 Nov 2020 00:00:00 +0000

Comparison of a genetic programming approach with ANFIS for power amplifier behavioral modeling and FPGA implementation

Mon, 01 Apr 2019 00:00:00 +0000

Machine Learning | Jose Alejandro Galaviz-Aguilar