SRAM-based FPGAs are increasingly being used in space applications. However, there are still many concerns about the reliability of these devices in high-radiation environments, particularly due to the possibility of single-event upsets (SEUs) in the configuration memory. This paper presents an architecture for implementing radiation-hardened SoCs based on FPGAs. Previous works used triple module redundancy (TMR) techniques together with scrubbing mechanisms based on partial reconfiguration. However, these solutions required external configuration controllers that increased the system complexity and deviated the design from the SoC principles. The proposed architecture uses novel self- reconfiguration techniques in order to eliminate the need for external components, so that a full radiation-hardened SoC can be implemented in a single FPGA. Since self- reconfiguration allows for on-board remote hardware updates, reliability is tackled at two key levels: Radiation- hardened operation and hardware upgradeability to solve design errors.
3 March 2011 in IEEE Transactions on Nuclear Science
Complexity of current digital systems and circuits involves new challenges in the field of hardening and measuring circuit's sensitivity under SEEs. In this work, a new solution for evaluating the SEU sensitivity of space systems based on using programmable logic devices is proposed. This solution is able to perform a deep analysis of fault effects in systems with hardware functionality distribution, taking into account the high complexity of the hardware nodes (complex programmable logic devices) and their collaborative hardening properties.
12 February 2019 in International Journal of Aerospace Engineering
The current evolution of the space missions demands to increase the computing capacities of the on-board computer while reducing its power consumption. This requirement evolves faster than the ability of the manufacturers to develop better space-qualified processors. To meet the strong requirements, the National Institute of Aerospace Technology has developed a distributed on-board computer based on commercial off-the-shelf (COTS). This computer, named OPTOS, provides enhanced computational capacities with respect to what computers of other small satellites typically provide. To maintain the reliability needed to perform typical critical activities such as real-time maintenance or current surveillance, authors have conceived a set of collaborative hardening techniques, taking advantage of the distributed architecture of the OPTOS On-Board Computer. The 3-year mission data analysis shows the feasibility of the collaborative hardening techniques implemented, despite using SEU sensitive devices. The authors describe the processes and tools used to analyse the data and clearly expose the functional errors found at unit level, while the system remains unfaulty and reliable thanks to the collaborative techniques.