A Review of Edge Intelligence Applications Based on RISC-V

Edge intelligence has emerged as a result of the merging of edge computing and artificial intelligence (AI) technologies. Nowadays, there is a lot of discussion on how to efficiently and cheaply execute AI applications on edge devices. In contrast to the traditional closed complex instruction set architecture (ISA), RISC-V is gradually being favored by edge devices due to its open source, scalability and flexibility. According to RISC-V’s capabilities, edge devices can design RISC-V extension modules on-demand for various intelligent application scenarios or leverage the low-power characteristics of RISC-V for core deployment. To our knowledge, there is a lack of a systematic introduction of RISC-V-based edge intelligent applications. The goal of this paper is to summarize RISC-V in edge intelligence applications from the hardware and software level. We focus on analyzing RISC-V-based accelerators, coprocessors, compilers and tool chains, as well as the main software applications in edge intelligence.