About / Short Bio

Prof. Fernanda Lima Kastensmidt is a Full Professor at the Federal University of Rio Grande do Sul (UFRGS), Brazil, where she leads internationally recognized research in dependable computing, semiconductor systems, and radiation-tolerant digital architectures. She received her B.Sc. in Electrical Engineering (1997), M.Sc. in Computer Science (1999), and Ph.D. in Computer Science (2003), all from UFRGS.

Her research lies at the intersection of computer architecture, microelectronics, artificial intelligence, and space systems, with a focus on designing reliable FPGA-based systems, RISC-V processors, AI accelerators, and heterogeneous System-on-Chip (SoC) platforms for operation in harsh environments. Her work advances methodologies for the qualification of digital systems under radiation through fault injection, heavy-ion and laser testing, reliability modeling, and hardware/software co-design. These technologies target aerospace, autonomous systems, safety-critical computing, and next-generation semiconductor platforms.

Prof. Kastensmidt has authored more than 300 scientific publications, co-authored several technical books, and wrote the Springer book Fault Tolerance Techniques for SRAM-Based FPGAs, considered one of the pioneering references in the field of FPGA reliability. Her contributions have helped establish new approaches for mitigating radiation-induced faults in programmable hardware and embedded computing systems.

She has played a leading role in the Brazilian space program, contributing to the development of scientific payloads for the NanoSatC-BR1 mission, launched in 2014, and the NanoSatC-BR2 mission, launched in 2021. These missions investigated the effects of the South Atlantic Anomaly (SAA) on modern integrated circuits, generating valuable in-orbit experimental data for the qualification of electronic systems used in space applications. She is currently leading the development of CubeLab, an educational CubeSat platform designed to validate artificial intelligence algorithms and Brazilian-designed integrated circuits in orbit.

Beyond her research, Prof. Kastensmidt has held several academic leadership positions at UFRGS, including Coordinator of the Graduate Program in Microelectronics (PGMICRO) and Head of the Department of Applied Informatics. She is also deeply committed to developing Brazil's semiconductor workforce.

She currently coordinates two of Brazil's largest national microelectronics education initiatives funded by the Brazilian Ministry of Science, Technology and Innovation (MCTI): CI-Innovator and CI-Expert. These programs provide advanced, industry-oriented training in integrated circuit design, verification, physical design, design-for-test (DFT), analog and mixed-signal circuits, embedded systems, and semiconductor technologies. Through strong partnerships with universities, research institutes, and semiconductor companies, the programs bridge academic education with practical engineering experience, preparing hundreds of undergraduate students, graduate students, and professionals to support the growth of Brazil's semiconductor ecosystem.

CI-Innovator (UFRGS): https://www.inf.ufrgs.br/inova-me/
CI-Expert (UFRGS): https://www.inf.ufrgs.br/ci-expert/

Today, Prof. Kastensmidt's research spans dependable AI hardware, RISC-V architectures, radiation-hardened computing, reconfigurable systems, chip qualification, and spaceborne computing, fostering innovation through close collaboration with academia, industry, and international research organizations. Her work combines fundamental research with technology transfer and workforce development, contributing to the advancement of resilient semiconductor technologies for future aerospace and critical computing applications.

Contact Info

E-mail

fglima at inf.ufrgs.br

Teaching Activities

Undergraduate course

01075

Arquitetura de Computadores - RISC-V

Undergraduate course

01175

Sistemas Digitais

Graduate course

MIC58

Teste e Confiabilidade de Sistemas de Hardware

Research Areas

Design of Fault Tolerant Systems: Processors, FPGAs, AI acelerators
Prof. Fernanda Lima Kastensmidt's research focuses on the design of dependable and fault-tolerant embedded computing systems for safety-critical applications, including aerospace and space systems. Her work spans the development of reliable FPGA-based architectures, RISC-V processors, AI accelerators, and heterogeneous System-on-Chip (SoC) platforms capable of operating under radiation-induced faults and harsh environments. Her research investigates hardware and software fault-tolerance techniques, radiation effects, hardware security, reliability-aware AI, and reconfigurable computing. She combines computer architecture, digital design, and machine learning to develop resilient computing platforms validated through fault injection, radiation testing, and in-orbit experimentation, bridging fundamental research with industrial and space applications.

Qualifying Digital Systems under Faults
My research focuses on the qualification of digital systems for operation in radiation environments encountered in space, high-altitude avionics, high-energy physics, and nuclear applications. The work encompasses the design, modeling, testing, and validation of radiation-tolerant FPGA-based systems, RISC-V processors, AI accelerators, and heterogeneous System-on-Chip (SoC) architectures. A major objective is to understand the impact of radiation-induced effects, such as Single Event Upsets (SEUs), Single Event Functional Interrupts (SEFIs), and Total Ionizing Dose (TID), on modern digital circuits and to develop cost-effective mitigation techniques that improve system reliability without excessive performance or power overhead. The research combines fault injection, laser testing, heavy-ion and proton irradiation experiments, reliability modeling, and hardware/software co-design to evaluate system robustness. The developed methodologies are validated on commercial programmable devices as well as custom integrated circuits, supporting the qualification of new digital technologies for aerospace and other safety-critical applications. Recent efforts also investigate the qualification of AI accelerators and RISC-V processors, enabling reliable onboard artificial intelligence and autonomous computing in future space missions.

SoC - Chiplets
Designing testing methodologies for chiplets based SoCs with AIB, UCie and other communication interfaces. IPs used: RISC-V, crossbars and AI accelerators.
Tapeouts in 65nm, 130nm, 180nm

Design of Cubesats
Design cubesats for education, research and qualification of chips developed in Brazil and algorithms with fault tolerant techniques.