arXiv subjects

Rafael Perez Martinez

Publications and source records attributed to Rafael Perez Martinez.

3 recordsLinked to original sources

arXiv2025

Robust Pareto Transistor Sizing of GaN HEMTs for Millimeter-Wave Applications

This paper introduces a robust Pareto design approach for transistor sizing of Gallium Nitride (GaN) High Electron Mobility Transistors (HEMTs), particularly for power amplifier (PA) and low-noise amplifier (LNA) designs in 5G applications. We consider five key design variables and two settings (PAs and LNAs) where we have multiple objectives. We assess designs based on three critical objectives, evaluating each by its worst-case performance across a range of Gate-Source Voltages ($V_{\text{GS}}$). We conduct simulations across a range of $V_{\text{GS}}$ values to ensure a thorough and robust analysis. For PAs, the optimization goals are to maximize the worst-case modulated average output power ($P_{\text{out,avg}}$) and power-added efficiency ($\text{PAE}_{\text{avg}}$) while minimizing the worst-case average junction temperature ($T_{\text{j,avg}}$) under a modulated 64-QAM signal stimulus. In contrast, for LNAs, the focus is on maximizing the worst-case maximum oscillation frequency ($f_{\text{max}}$) and Gain, and minimizing the worst-case minimum noise figure ($\text{NF}_{\text{min}}$). We utilize a derivative-free optimization method to effectively identify robust Pareto optimal device designs. This approach enhances our comprehension of the trade-off space, facilitating more informed decision-making. Furthermore, this method is general across different applications. Although it does not guarantee a globally optimal design, we demonstrate its effectiveness in GaN transistor sizing. The primary advantage of this method is that it enables the attainment of near-optimal or even optimal designs with just a fraction of the simulations required for an exhaustive full-grid search.

eess.SY↗

arXiv2024

Compact Model Parameter Extraction via Derivative-Free Optimization

In this paper, we address the problem of compact model parameter extraction to simultaneously extract tens of parameters via derivative-free optimization. Traditionally, parameter extraction is performed manually by dividing the complete set of parameters into smaller subsets, each targeting different operational regions of the device, a process that can take several days or weeks. Our approach streamlines this process by employing derivative-free optimization to identify a good parameter set that best fits the compact model without performing an exhaustive number of simulations. We further enhance the optimization process to address three critical issues in device modeling by carefully choosing a loss function that focuses on relative errors rather than absolute errors to ensure consistent performance across different orders of magnitude, prioritizes accuracy in key operational regions above a specific threshold, and reduces sensitivity to outliers. Furthermore, we utilize the concept of train-test split to assess the model fit and avoid overfitting. We demonstrate the effectiveness of our approach by successfully modeling a diamond Schottky diode with the SPICE diode model and a GaN-on-SiC HEMT with the ASM-HEMT model. For the latter, which involves extracting 35 parameters for the ASM-HEMT DC model, we identified the best set of parameters in under 6,000 trials. Additional examples using both devices are provided to demonstrate robustness to outliers, showing that an excellent fit is achieved even with over 25% of the data purposely corrupted. These examples demonstrate the practicality of our approach, highlighting the benefits of derivative-free optimization in device modeling.

cs.LG↗

arXiv2022

Visual anomaly detection in video by variational autoencoder

Video anomalies detection is the intersection of anomaly detection and visual intelligence. It has commercial applications in surveillance, security, self-driving cars and crop monitoring. Videos can capture a variety of anomalies. Due to efforts needed to label training data, unsupervised approaches to train anomaly detection models for videos is more practical An autoencoder is a neural network that is trained to recreate its input using latent representation of input also called a bottleneck layer. Variational autoencoder uses distribution (mean and variance) as compared to latent vector as bottleneck layer and can have better regularization effect. In this paper we have demonstrated comparison between performance of convolutional LSTM versus a variation convolutional LSTM autoencoder

cs.CV↗