Skip to main content

Resultados relacionados

Resultados totales (Incluyendo duplicados): 4
Encontrada(s) 1 página(s)

DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Otros (Other)

AISAD, Artificial Intelligence-assisted Synthesis toolbox for Analog/Digital interfaces

  • Liñán-Cembrano, Gustavo
  • Rosa, José M. de la
Version: v1.0. -- Copyright (c) 2025 IMSE-CNM (CSIC-Univ. of Seville) AI-Supported Microelectronic Design Group. Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. The software is provided "as is", without warranty of any kind, express or implied, including but not limited to the warranties of merchantability, fitness for a particular purpose and noninfringement. in no event shall the authors or copyright holders be liable for any claim, damages or other liability, whether in an action of contract, tort or otherwise, arising from, out of or in connection with the software or the use or other dealings in the software., AISAD (Artificial Intelligence-assisted Synthesis toolbox for Analog/Digital interfaces) is a MATLAB-based software toolkit enhanced with artificial neural networks for the design of Sigma-Delta Modulators (SDMs). It supports designers in finding suitable architectures and parameters that meet given performance specifications. AISAD integrates a Graphical User Interface (GUI) and allows inverse regression and classification-based synthesis, enabling both novice and expert users to perform design space exploration and validation through simulation. It includes support for behavioral validation via SIMSIDES. Noteworthy features include automated model training pipelines, inverse prediction (from specs to design parameters), and behavioral validation of results., This work was supported in part by Grants PID2022-138078OB-I00, and PDC2023-145808-I00, funded by MCIN/AEI/10.13039/501100011033,Ministry for Digital Transformation and Public Function through grant USECHIP (TSI-069100-2023-001) of PERTE Chip Chair program, by the European Union NextGenerationEU/PRTR, by ERDF A way of making Europe, and by CSIC AI-Hub initiative (CSIC Interdisciplinary Thematic Platform on Artificial Intelligence)., File List: - `PROCESSING/`: scripts for dataset preparation and visualization. - `CLASSIFIERS/`: Python modules for model training and evaluation. - `REGRESSION-ANN/`: TensorFlow/Keras scripts for regression. - `VALIDATION/`: modules for behavioral validation and cross-validation. - `GUI/`: MATLAB files for the graphical interface. - `MODELS/`: Simulink models used in simulations. - `SPECS_PREDICTOR_ANN/`: tools for inverse prediction. - `requirements.txt`: list of required Python packages., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Addressing preprocessing for spectrum sensing using image processing

  • Rojas, Andrés
  • Jovanovic-Dolecek, G.
  • Rosa, José M. de la
This paper presents a novel approach to preprocessing spectrograms for spectrum sensing (SS) from the image-processing perspective. Gaussian bilateral filtering, a common image-denoising technique, has been introduced to improve spectrograms in SS in high-noise environments. This approach is evaluated by simulating LTE and 5 G NR signal spectrograms across various signal-to-noise ratios (SNRs). An extensive review and comparison with recent spectrogram-based works for different applications demonstrated that the proposed approach does not depend on deep learning models to denoise spectrograms, showing a simpler yet effective strategy to address SS., This work was supported in part by Consejo Nacional de Humanidades, Ciencias y Tecnologías (CONAHCYT) and by Grants PID2019–103876RB-I00, PID2022-138078OB-I00, PDC2023-145808-I00 funded by MCIN/AEI/10.13039/501100011033 and by the European Union, ERDF A way of making Europe., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Artículo científico (JournalArticle)

Prediction-Based Spectrum Sensing Framework for Cognitive Radio

  • Rojas, Andrés
  • Follet, Gawen
  • Jovanovic-Dolecek, G.
  • Rosa, José M. de la
  • Liñán-Cembrano, Gustavo
This paper presents a hardware-software deep learning architecture for prediction-based spectrum sensing in Cognitive Radio (CR) applications. A convolutional neural network-based predictor for spectrum occupancy was trained using the band power from I/Q samples acquired by a softwaredefined radio (SDR). Additionally, a second neural engine was trained for radio frequency (RF) frame detection based on spectrograms. We implemented a transfer-learning solution using a You-Only-LookOnce version 8 nano model with a synthetic dataset comprising thousands of wireless signals, including Wi-Fi, Bluetooth, and collision frames. Once trained, the two neural networks were transferred to a Raspberry Pi 5 - an affordable single-board computer - connected to two (one for Rx, one for Tx) ADALM-PLUTO SDR systems for benchmarking using over-the-air signals in the 2.4 GHz band. Together with our methodology and experimental results, the paper also presents an overview of current spectrum prediction proposals and RF frame detectors. Remarkably, to the best of our knowledge, this proposed framework is the first approach towards an Internet of Things-suitable implementation of prediction-based spectrum sensing for CR applications., This publication has been funded by grant USECHIP (TSI-069100-2023-001), project funded by the Secretary of State for Telecommunications and Digital Infrastructure, Ministry for the Digital Transformation and Civil Service and by the European Union Next Generation EU/PRTR. This work was also supported in part by Secretaría de Ciencia, Humanidades, Tecnología e Innovación (Secihti) and by Grants PID2022-138078OB-I00, and PDC2023-145808-I00, funded by MICIU/AEI/10.13039/501100011033 and by the European Union, ERDF A way of making Europe., Peer reviewed
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publication Otros (Other)

Optimized Design of 𝚺𝚫 Modulators Using Deep-Learningand Simulated Annealing

  • Liñán-Cembrano, Gustavo
  • Rosa, José M. de la
The data that support the findings of this study will be openly available at https://git.csic.es/27317787C/aisad. Once the paper will be accepted, the following data will be provided a doi and repository name will be provided., This letter presents a MATLAB toolbox for the automated high-level design and optimization of analogue-to-digital converters (ADCs), using sigma-delta modulators ((Formula presented.)) as case studies. The tool combines machine learning (ML) techniques and behavioural simulation to obtain the optimum set of building-block (amplifiers, comparators, etc.) requirements for a given set of specifications, namely resolution, signal bandwidth and power consumption. Two machine learning blocksgradient boosting classifiers and regression-type artificial neural networks—are trained, using Python libraries, to identify the best ADC architecture as well as to infer a set of design parameters which yields ADC specifications. The result from the ML blocks can be cross-checked in behavioural simulations in MATLAB and also optimized with respect to signal-to-noise ratio (SNR), power consumption, or figure-of-merit (FoM) using an embedded simulated annealing (SA) process. The toolbox is controlled through a graphical user interface (GUI) for MATLAB which guides the designer through the whole process, from specifications to obtaining an implementation that meets the required specifications., This work was supported in part by Grants PID2022-138078OB-I00, and PDC2023-145808-I00, funded by MCIN/AEI/10.13039/501100011033, and by Ministry for Digital Transformation and Public Function through grant USECHIP (TSI-069100-2023-001) of PERTE Chip Chair program, by the European Union NextGenerationEU/PRTR, by ERDF a way of making Europe., Peer reviewed