Journal of Biomedical and Engineering Research
ISSN: 3065-8780 | DOI: 10.64030/3065-8780
Rajesh Kumar Singh
Modern control systems increasingly rely on rapid, reliable, and adaptable signal processing to maintain operational stability, particularly in environments where conditions change quickly. To address these requirements, this work introduces a twodimensional finite-impulse-response (2-D FIR) filter framework designed specifically for Field-Programmable Gate Arrays (FPGAs). The architecture combines a memory-centric filtering approach with a compact machine-learning module capable of anticipating instability in real time. Distributed arithmetic is employed to reduce computational overhead, while a streamlined support-vector-machine classifier evaluates spatial-temporal features derived from the filtered data. When implemented on a Xilinx Artix-7 device, the proposed system demonstrated a noticeable improvement in throughput, reduced power consumption, and a high level of prediction accuracy. These results highlight the framework’s potential as a practical and scalable solution for embedded applications that depend on fast, deterministic processing and early stability assessment.