Rodolfo S. Villaca
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Rodolfo S. Villaca has a total of 125 co-authors. Here are the top co-authors by number of shared publications:
- Vinicius Fernandes Soares Mota
Universidade Federal do Espírito Santo (UFES)
7 shared publications - Johann Jakob Schmitz Bastos
UFES
5 shared publications - João Pedro Camargo Batista
UFES
5 shared publications - Cristina Klippel Dominicini
IFES
4 shared publications - Moises Renato Nunes Ribeiro
UFES
4 shared publications - Eduardo Cerqueira
Federal University of Pará, Belém, Pará, Brazil
3 shared publications
Latest Publications
- Open Access
- 24 July 2026
- Published by Elsevier in Journal of Network and Computer Applications
- Vol. 254
Secure and efficient Federated Learning using sketches and Fully Homomorphic Encryption
- 24 July 2026
- Published by Elsevier in Journal of Network and Computer Applications
- Vol. 254
Open AccessSecure and efficient Federated Learning using sketches and Fully Homomorphic Encryption
- Open Access
- 19 July 2026
- Published by Sociedade Brasileira de Computação
- p. 388-399
Implementation and Analysis of a Low-Power Heart Rate Prediction Solution Using Hardware Acceleration on the ESP32-S3 Microcontroller
This work details the implementation of an accelerometry-based heart-rate estimation system optimized for the ESP32-S3 microcontroller. It leverages PIE-based Single Instruction, Multiple Data (SIMD) extensions via ESP-DSP library to accelerate the pre-processing, training and inference of a continual-learning model. Under a modular FreeRTOS architecture, the system achieved 4.99 µs training latency, requiring less than 1 kB of RAM for the ML model itself. Results show a 30.6% computational performance gain and 8% energy consumption reduction with acceleration enabled. These efficiencies allow race-to-sleep strategies that extend the estimated battery life to 51 days on COTS hardware.
- 19 July 2026
- Published by Sociedade Brasileira de Computação
- p. 388-399
Open AccessImplementation and Analysis of a Low-Power Heart Rate Prediction Solution Using Hardware Acceleration on the ESP32-S3 Microcontroller
This work details the implementation of an accelerometry-based heart-rate estimation system optimized for the ESP32-S3 microcontroller. It leverages PIE-based Single Instruction, Multiple Data (SIMD) extensions via ESP-DSP library to accelerate the pre-processing, training and inference of a continual-learning model. Under a modular FreeRTOS architecture, the system achieved 4.99 µs training latency, requiring less than 1 kB of RAM for the ML model itself. Results show a 30.6% computational performance gain and 8% energy consumption reduction with acceleration enabled. These efficiencies allow race-to-sleep strategies that extend the estimated battery life to 51 days on COTS hardware.
- 25 May 2026
- Published by Sociedade Brasileira de Computação
- p. 105-113
MedTracker: A BLE-Based Indoor Localization System for Tracking Portable Medical Devices in Hospital Environments
The misplacement and inefficient tracking of Portable Medical Devices (PMDs) represent a persistent operational challenge in hospital environments, leading to delays in care delivery and increased burden on medical staff. This paper presents MedTracker, an indoor localization tool based on Bluetooth Low Energy (BLE), designed for real-time asset tracking in healthcare facilities. The system employs BLE tags attached to monitored devices and strategically deployed ESP32-based gateways that forward signal data to a central server. By combining BLE signals and filtering techniques with machine learning classification and clustering algorithms, MedTracker is capable of inferring room- and floor-level device location, as well as detecting location changes over time. The proposed tool was motivated by practical demands observed at a public university hospital and validated through experimental evaluation, demonstrating its effectiveness as a scalable and deployable solution for hospital asset management.
- 25 May 2026
- Published by Sociedade Brasileira de Computação
- p. 105-113
MedTracker: A BLE-Based Indoor Localization System for Tracking Portable Medical Devices in Hospital Environments
The misplacement and inefficient tracking of Portable Medical Devices (PMDs) represent a persistent operational challenge in hospital environments, leading to delays in care delivery and increased burden on medical staff. This paper presents MedTracker, an indoor localization tool based on Bluetooth Low Energy (BLE), designed for real-time asset tracking in healthcare facilities. The system employs BLE tags attached to monitored devices and strategically deployed ESP32-based gateways that forward signal data to a central server. By combining BLE signals and filtering techniques with machine learning classification and clustering algorithms, MedTracker is capable of inferring room- and floor-level device location, as well as detecting location changes over time. The proposed tool was motivated by practical demands observed at a public university hospital and validated through experimental evaluation, demonstrating its effectiveness as a scalable and deployable solution for hospital asset management.
- Open Access
- 8 April 2026
- Published by Elsevier in Computer Networks
- Vol. 282
Building realistic cybersecurity datasets through a edge cloud testbed with cost-aware monitoring
- 8 April 2026
- Published by Elsevier in Computer Networks
- Vol. 282
Open AccessBuilding realistic cybersecurity datasets through a edge cloud testbed with cost-aware monitoring
- Open Access
- 4 April 2026
- Published by Elsevier in SSRN Electronic Journal
Secure and Efficient Federated Learning Using Sketches and Fully Homomorphic Encryption
Federated Learning (FL) represents a distributed training paradigm designed to enable collaborative model training without sharing raw data. However, sophisticated attackers can infer or reconstruct sensitive client data from shared model updates, thereby weakening privacy protection for clients. Moreover, FL also struggles with the data heterogeneity (non-IID distribution) among participants and with high communication overhead due to the transmission of large model weights over multiple rounds. This work presents an integrated framework of three novel algorithms to address the aforementioned issues in FL: i) FedSketch leverages Differential Privacy (DP) with probabilistic count sketches to compress high-dimensional model updates. This compression approach reduces communication overhead while enhancing privacy. ii) The CKKSFED algorithm, which utilizes Fully Homomorphic Encryption (FHE) to perform client clustering, mitigating the negative effects of non-IID data distribution in a privacy-preserving manner. iii) The MetricBasedSelection algorithm selects clients in each round based on metrics to improve convergence and reduce communication. We evaluate the effectiveness of our solutions in an emulation environment using image classification and real-world time series datasets. FedSketch significantly reduced communication overhead, achieving up to a 73-fold reduction in model size. Clustering clients with CKKSFED achieved convergence in approximately half the number of rounds compared to non-clustered methods in non-IID scenarios. Finally, FedSketch, with the metric-based selection algorithm, selected fewer clients for training in each round, thereby contributing to the reduction of communication costs. Jointly, these algorithms reduced communication overhead while ensuring strong privacy guarantees.
- 4 April 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessSecure and Efficient Federated Learning Using Sketches and Fully Homomorphic Encryption
Federated Learning (FL) represents a distributed training paradigm designed to enable collaborative model training without sharing raw data. However, sophisticated attackers can infer or reconstruct sensitive client data from shared model updates, thereby weakening privacy protection for clients. Moreover, FL also struggles with the data heterogeneity (non-IID distribution) among participants and with high communication overhead due to the transmission of large model weights over multiple rounds. This work presents an integrated framework of three novel algorithms to address the aforementioned issues in FL: i) FedSketch leverages Differential Privacy (DP) with probabilistic count sketches to compress high-dimensional model updates. This compression approach reduces communication overhead while enhancing privacy. ii) The CKKSFED algorithm, which utilizes Fully Homomorphic Encryption (FHE) to perform client clustering, mitigating the negative effects of non-IID data distribution in a privacy-preserving manner. iii) The MetricBasedSelection algorithm selects clients in each round based on metrics to improve convergence and reduce communication. We evaluate the effectiveness of our solutions in an emulation environment using image classification and real-world time series datasets. FedSketch significantly reduced communication overhead, achieving up to a 73-fold reduction in model size. Clustering clients with CKKSFED achieved convergence in approximately half the number of rounds compared to non-clustered methods in non-IID scenarios. Finally, FedSketch, with the metric-based selection algorithm, selected fewer clients for training in each round, thereby contributing to the reduction of communication costs. Jointly, these algorithms reduced communication overhead while ensuring strong privacy guarantees.
- 18 December 2025
- Published by Springer Nature in Annals of Telecommunications
- Vol. 81 (5-6) , 429-450
PORVIR-5G: Programmability, ORchestration, and VIRtualization of next-generation networks
- 18 December 2025
- Published by Springer Nature in Annals of Telecommunications
- Vol. 81 (5-6) , 429-450
PORVIR-5G: Programmability, ORchestration, and VIRtualization of next-generation networks
- Open Access
- 8 September 2025
- Published by Association for Computing Machinery (ACM)
- p. 130-131
A Lightweight Emulation Framework for Energy-Aware Federated Learning
- 8 September 2025
- Published by Association for Computing Machinery (ACM)
- p. 130-131
Open AccessA Lightweight Emulation Framework for Energy-Aware Federated Learning
- 29 June 2025
- Published by Springer Nature
- p. 345-360
A Body Tracking Framework for Neuromotor Rehabilitation with Multimedia Application Support
- 29 June 2025
- Published by Springer Nature
- p. 345-360
A Body Tracking Framework for Neuromotor Rehabilitation with Multimedia Application Support
- 28 June 2025
- Published by Springer Nature
- p. 65-80
Correlating Socioeconomic Factors and Syphilis Incidence: A Case Study in Brazil
- 28 June 2025
- Published by Springer Nature
- p. 65-80
Correlating Socioeconomic Factors and Syphilis Incidence: A Case Study in Brazil
- 19 May 2025
- Published by Sociedade Brasileira de Computação
- p. 70-83
Gerenciamento de Energia no Mercado Livre com Usinas Virtuais e Contratos Inteligentes via Blockchain
Este artigo propõe um sistema de gerenciamento de energia no mercado livre que utiliza blockchain e contratos inteligentes para intermediar a interação entre prosumidores, superando as limitações do Sistema de Compensação de Energia Elétrica (SCEE) — que atualmente impede a transferência de créditos excedentes. A solução integra Virtual Power Plants (VPP), utilizando tokens como representação de créditos energéticos, com um contrato inteligente que gerencia automaticamente sua compra e venda. O sistema foi simulado com dados reais de consumo e geração fotovoltaica. A implementação utiliza a blockchain Sepolia e oráculos Chainlink para obter as tarifas de energia usadas pelas concessionárias. Os resultados mostram que o sistema proporciona aos usuários: autonomia, redução de custos e uma gestão eficiente dos recursos energéticos.
- 19 May 2025
- Published by Sociedade Brasileira de Computação
- p. 70-83
Gerenciamento de Energia no Mercado Livre com Usinas Virtuais e Contratos Inteligentes via Blockchain
Este artigo propõe um sistema de gerenciamento de energia no mercado livre que utiliza blockchain e contratos inteligentes para intermediar a interação entre prosumidores, superando as limitações do Sistema de Compensação de Energia Elétrica (SCEE) — que atualmente impede a transferência de créditos excedentes. A solução integra Virtual Power Plants (VPP), utilizando tokens como representação de créditos energéticos, com um contrato inteligente que gerencia automaticamente sua compra e venda. O sistema foi simulado com dados reais de consumo e geração fotovoltaica. A implementação utiliza a blockchain Sepolia e oráculos Chainlink para obter as tarifas de energia usadas pelas concessionárias. Os resultados mostram que o sistema proporciona aos usuários: autonomia, redução de custos e uma gestão eficiente dos recursos energéticos.