FPGA Platforms and Their Use in Edge Computing

Abstract

The article examines the role and future prospects of programmable logic devices (FPGAs) and system-on-chip FPGA (SoC FPGA) platforms within the edge computing paradigm. Particular attention is given to the combination of adaptability, fine-grained power consumption control, and high degrees of parallelism, which are critical characteristics for modern edge platforms. Additionally, the current state of FPGA adoption in practical edge scenarios is analyzed, ranging from video analytics in transportation systems to industrial vibration diagnostics and acceleration of telecommunications functions. Examples of both conventional FPGA-based solutions and hybrid SoC FPGA architectures are discussed, where programmable logic is tightly integrated with ARM-based processors to achieve balanced workload distribution between software and hardware components. It is demonstrated that such systems can significantly reduce processing latency, optimize energy consumption, and enable autonomous operation even under remote or unstable network conditions. The role of embedded operating systems is also examined, particularly in the context of SoC FPGA platforms, where Linux-based environments enable separation between control logic and hardware-accelerated data paths. The paper discusses how embedded operating systems influence system flexibility, software reuse, update mechanisms, and long-term maintainability of FPGA-based edge devices. In addition, the work addresses technical and organizational barriers that limit large-scale deployment of FPGA-based edge systems, including the lack of unified high-level synthesis toolchains, the steep learning curve associated with HDL-based design, and limited support for infrastructure-oriented workflows in FPGA-centric environments. Approaches for dynamic logic reconfiguration aimed at improving adaptability of local IoT systems are considered, along with challenges related to logic testing, system configuration, and scalability when adapting platforms to new application requirements. Modern development tools and frameworks for flexible system design, including cloud-based services and high-level programming environments such as Vitis HLS and Intel oneAPI, are also discussed in the context of reducing development complexity and accelerating design iterations.