Engenharia Elétrica

URI Permanente desta comunidade

https://repositorio.fei.edu.br/handle/FEI/21

Navegar

Navegando Engenharia Elétrica por Autor "ADDED, N."

Agora exibindo 1 - 4 de 4
  • Nenhuma Miniatura disponível
    Artigo de evento
    Brazilian facilities to study radiation effects in electronic devices
    (2013-09-27) MEDINA, N.H.; Marcilei Aparecida Guazzelli; ADDED, N.; AGUIAR, V. A. P.; AGUIRRE, F.; Renato Giacomini; MACCHIONE, E. L. A.; DE MELO, M. A. A.; OLIVEIRA, J. A.; Roberto Santos; SEIXAS, L. E.; TABACNIKS, M. H.
    © 2013 IEEE.Three facilities in Brazil are being prepared and upgraded to test and to qualify electronic devices regarding their tolerance to TID and SEE: a 60 kV X-ray source, a 1.7 MV Pelletron accelerator for low energy proton beams and an 8 MV Pelletron accelerator that produces heavy ion beams. MOSFET transistors were exposed to 10-keV X-rays and to 2.4 MeV protons extracted into air. During irradiation, characteristic curves were continuously measured to monitor the circuit's behavior relative to the accumulated dose. 12C, 16O, 28Si, 35Cl and 63Cu heavy ion beams were also used mostly to test the experimental setup, and verify beam uniformity at low fluence conditions, equilibrium charge state, and carbon stripper foil durability. To test the setup for SEE, a pMOS transistor was irradiated with 63 MeV 63Cu ions scattered at 15° by a 275 μg/cm2 gold foil. The setups are now available for TID and SEE studies in electronic devices.
  • Nenhuma Miniatura disponível
    Artigo de evento
    Heavy Ions Testing of an All-Convolutional Neural Network for Image Classification Evolved by Genetic Algorithms and Implemented on SRAM-Based FPGA
    (2019-09-20) BENEVENUTI, F.; DE OLIVEIRA, A. B.; LOPES, I. C.; KASTENSMIDT, F. L.; ADDED, N.; AGUIAR, V. A P.; MEDINA, N.H.; Marcilei Aparecida Guazzelli
    This work investigates the vulnerability of an image classification engine under heavy-ions accelerated irradiation. The engine is based on all-convolutional neural-network trained with the GTSRB traffic sign recognition benchmark and embedded into 28nm SRAM-based FPGA.
  • Nenhuma Miniatura disponível
    Artigo
    Lockstep Dual-Core ARM A9: Implementation and Resilience Analysis under Heavy Ion-Induced Soft Errors
    (2018-08-05) DE OLIVEIRA, A. B.; RODRIGUES, G. S.; KASTENSMDT, F. L.; ADDED, N.; MACCHIONE, E. L. A.; AGUIAR, V . A. P.; MEDINA N. H.; Salvador Gimenez
    © 1963-2012 IEEE.This paper presents a dual-core lockstep (DCLS) implementation to protect hard-core processors against radiation-induced soft errors. The proposed DCLS is applied to an Advanced RISC Machine Cortex-A9 embedded processor. Different software optimizations were evaluated to assess their impact on performance and fault tolerance. Heavy ions' experiments and fault injection emulation were performed to analyze the system susceptibility to errors and the DCLS performance. Results show that the approach is able to decrease the system cross section and achieve high protection against errors. The DCLS successfully protects the system from up to 78% of the injected faults. The execution performance analysis shows that by reducing the number of verifications and augmenting the block partition execution time, it is possible to increase the system reliability with minimal performance losses.
  • Nenhuma Miniatura disponível
    Artigo de evento
    Reliability Evaluation of Voters for Fault Tolerant Approximate Systems
    (2021-10-27) BALEN, T. R.; GONZALEZ, C. J.; OLIVEIRA, I. F. V.; SCHVITTZ, R. B.; ADDED, N.; MACCHIONE, E. L. A.; AGUIAR, V. A. P.; Marcilei Aparecida Guazzelli; MEDINA, N. H.; BUTZEN, P. F.
    This work presents a study on the reliability of voters for approximate fault tolerant systems in the context of single event effects. A first case study analyses different topologies of single-bit majority voters for logic circuits by means of fault injection by simulation. In these simulations a previous analysis is performed identifying the critical diffusion areas of the physical implementation according to the voter input vector. Additionally, as second case study, practical heavy ion experiments on different architectures of software-based approximate voters for mixed-signal applications are also presented, and the cross section of each voter is evaluated. The system comprising the voter was irradiated in two distinct experiments with an 16O ion beam, producing an effective $LET$ at the active region of 5.5 MeV/mg/cm2. Results of both case-studies allow identifying the most tolerant voter architectures (among the studied ones) for approximate computing applications under single event effects.