Implementation of a reconfigurable computing system for space applications

Abstract

High reliability is usually the most important requirement in space applications. There is no hardware repair for them. Long mission times and harsh environment are a challenge for electronic circuits, and particular error mitigation techniques have to be implemented in order to be able to cope with the expected error effects. Always, Triple Modular Redundancy (TMR) is relied mostly to prevent SEUs from causing functional errors. And in recent decades, SRAM-based FPGAs are used a lot in space for its high performance and flexibility. However, use of FPGAs in space applications can also be challenging. They are too susceptible to transient faults, such as SEUs. Reconfiguration provides more possibilities for SRAM-based FPGAs into space. Reconfigurable technology solves the disadvantage that the TMR modules cannot heal the error themselves. Additionally, reconfiguration is an improvement in terms of resource utilization and costs. In this paper, TMR and reconfigurable computing are combined to level up the reliability of the space applications. The paper proposed a reconfigurable computing architecture based on a TMR system. A new kind of reconfigurable architecture using SoP (system-on-a-package) technology is presented. Also, a whole space application will be presented.