Redundant data types and operations in HLS and their use for a robot controller unit fault tolerance evaluation

Abstract

Some environments (e.g. space, aerospace or medical systems) require electronic systems to withstand an increased, occurrence of faults. Moreover, the failure of these electronic systems can cause high economical losses or endanger human health. Fault tolerance is one of the techniques, the goal of which is to avoid such situations. This paper presents an approach to evaluate the degree of importance of individual system partitions when High-Level Synthesis (HLS) methodology is used. The importance of individual partitions was evaluated by the usage of our approach to fault-tolerant data-paths design which is based on the HLS input specification modification. The partitions are formed by sets of variables and operations. A brief description of the approach to fault tolerance in HLS is shown in the paper as well. Our experiments are evaluated using an SRAM-based FPGA evaluation platform which allows us to analyze fault tolerance properties of the Design Under Test (DUT). In the evaluation platform, functional verification in combination with fault injection is utilized.