Majority Type and Redundancy Level Influences on Redundant Data Types Approach for HLS

conference paper

English

Due to the increasing demand for reliable computation in environments that require electronic systems to withstand an increased occurrence of faults (e.g. space, aerospace and medicine), new techniques of the so-called Fault Tolerance insertion arise. From another perspective, today's systems have become incredibly large and complex. New methodologies (e.g. High-Level Synthesis) are used to reduce time to market and simplify the verification of the resulting system. In our research we focus on an implementation of Fault Tolerance into complex systems with the usage of High-Level Synthesis. In our approach, we are using newly designed Data Types that introduce redundancy on the functional level of an algorithm. In this student paper, our previously presented technique is extended by another means of redundancy and also by a new type of voting component. The systems incorporating various levels of redundancies using our approach are experimentally tested on the application of a robot controller. The paper also briefly presents the evaluation process and investigates its correct settings. The results show that the bit-based majority function is more suitable for usage with our Redundant Data Types.

High-Level Synthesis, Redundant Data Type, Level of Redundancy, Voter Component, CatapultC, Fault Tolerance.

LOJDA, J.; PODIVÍNSKÝ, J.; KOTÁSEK, Z.; KRČMA, M.

8. 10. 2018

IEEE Computer Society

Tallinn

978-1-5386-7312-6

2018 16th Biennial Baltic Electronics Conference (BEC)

1

4

4

https://www.fit.vut.cz/research/publication/11600/