Radiation tolerant many-core computing system for aerospace applications

Abstract

When integrated circuits are exposed to ionizing radiation, a variety of fault conditions can occur. This draws concern to the aerospace community as they look toward integrating more complex computing systems into flight applications. The detrimental effects that radiation can have on integrated circuits can be broken up into two categories: single event effects and total ionizing dose. Single event effects refer to non-destructive electron hole pairs that are created by the radiation which can lead to logical failures. Total ionizing dose refers to the permanent damage to a device caused by the electron hole pairs getting trapped prior to recombination and results in oxide breakdown and leakage current. In order to provide a robust computing platform for aerospace applications, both of these effects must be addressed. This thesis presents a set of novel fault mitigation strategies to increase the reliability of aerospace flight computers by exploiting the reconfigurability of Field Programmable Gate Arrays. First, redundant circuitry and a voting system are used to recover from non-destructive faults. Secondly, spare circuitry is used to spatially avoid faults and replace permanently damaged circuitry. Finally, partial reconfiguration of the Field Programmable Gate Array is used to repair faults in the configuration memory of the device. These fault mitigation techniques are all combined into a complete system to provide a robust computing platform for aerospace applications.