Hybrid Technique for Soft Error Detection in Dependable Embedded Software: a First Experiment

Abstract

Embedded systems’ hardware can be impacted by soft errors, which causes either a data flow error or a control flow error in the systems’ software. To counter such errors, numerous software-implemented techniques have been proposed to detect either one of them. However, there exist few techniques that are designed to detect both types of errors. This paper aims to fill that gap by proposing a software-implemented technique that has been designed to detect both data flow and control flow errors, called Data and Control Flow Error Detection (DCFED). We verified the technique using a fault injection campaign and compared the measured results to those of a similar technique, called Software Implemented Error Detection (SIED). The results show that DCFED achieves a higher error detection ratio.