A Low-Power FPGA-Based Edge Computing Architecture for Real-Time Smart Traffic Monitoring and Vehicle Classification

Abstract

The rapid growth of urbanization and vehicular transportation has significantly increased traffic congestion, road accidents, fuel consumption, and environmental pollution across modern smart cities. Conventional cloud-based intelligent transportation systems rely heavily on centralized processing, which introduces communication latency, high bandwidth consumption, privacy concerns, and increased energy requirements, making them unsuitable for real-time traffic monitoring applications. Edge computing has emerged as a promising paradigm by enabling data processing closer to the data source, thereby reducing latency and network dependency. Simultaneously, Field Programmable Gate Arrays (FPGAs) have gained considerable attention for edge Artificial Intelligence (AI) applications because of their parallel processing capability, hardware reconfigurability, low power consumption, and real-time computational efficiency. This research presents the development of a low-power FPGA-based edge computing architecture for real-time smart traffic monitoring and vehicle classification. The proposed architecture integrates high-definition image acquisition, FPGA-based image preprocessing, feature extraction, lightweight convolutional neural network (CNN) inference, vehicle classification, traffic density estimation, and intelligent decision-making within a single edge platform. The architecture is implemented using a Xilinx FPGA integrated with an embedded ARM processor to achieve high-speed processing while minimizing power consumption. Hardware performance is evaluated using latency analysis, resource utilization, throughput, classification accuracy, and power measurements. Experimental results demonstrate that the proposed FPGA-based architecture achieves high vehicle classification accuracy, ultra-low inference latency, reduced energy consumption, and efficient real-time operation under diverse traffic conditions. The developed system provides an effective solution for smart traffic management, intelligent transportation systems, autonomous vehicles, adaptive traffic signal control, and future smart city infrastructure.