Reconfigurable Computing for Next-Generation Embedded Systems: A Comprehensive Survey of Architectures, Frameworks, and Applications

Abstract

Reconfigurable computing (RC) has recently become a very important paradigm in scaling up performance, flexibility and energy efficiency of next generation embedded systems. Fixed-function microcontrollers and general-purpose processors, the kinds of traditional embedded platforms, are sometimes unable to suit the demands of modern applications in the Internet of Things (IoT), artificial intelligence (AI), automotive electronics, aerospace control systems, and real-time signal processing, due to their strictness and dynamism. On the contrary, RC technologies, and especially the Field-Programmable Gate Arrays (FPGAs), Coarse-Grained Reconfigurable Arrays (CGRAs), and dynamic partial reconfiguration (DPR) enable creating such a solution in the form of run-time reshaping of hardware resources to fit the changing workloads, without going through a redesign or performing power-thirsty overprovisioning. The survey is a comprehensive review of reconfigurable architecture development path, starting with the conventional RTL-based FPGA-based implementation to high-level synthesis (HLS)-based and AI-accelerated design frameworks, where development time is greatly decreased at the expense of no reduction in performance optimization. We review a taxonomy of the hardware / software co-design methodologies, reconfiguration strategies and abstraction frameworks and compare their performance with benchmarks and real world deployment. Moreover, we give a detailed analysis of application-specific recipes where RC has had a transformative effect such as energy-efficient machine learning inference at edge, cyber-physical system adaptation at runtime, and reconfigurable cryptographic primitives on secure embedded systems. Research issues considered as critical and critical research challenges that need to be resolved in the paper include the complexity of toolchains, power-aware task scheduling, bitstream security, and hardware-software interoperability which are essential to get the maximum out of RC. The future trend of the field is indicated as well, with emerging trends like AI-driven reconfiguration orchestration, RISC-V based reconfigurable SoCs and 3D reconfigurable architectures. Integrating architectural advancements, structure growth, and implementation use in areas of reconfigurable embedded computing, this extended survey is beneficial to researcher and practitioner in developments to promote the advancement of reconfigurable embedded computing.