Traffic congestion remains a major challenge in urban areas due to the high cost, scalability issues, and inefficiencies of traditional monitoring systems. This study proposes an innovative, cost-effective traffic monitoring system utilizing Android-based Smart Roadside Units (RSUs) to detect vehicles and analyze real-time traffic data. The system leverages the You Only Look Once, version 8 (YOLOv8) model, enhanced with the Slicing Aided Hyper Inference (SAHI) algorithm to improve detection accuracy for small and distant objects. Field experiments were conducted using three Android device categories high, medium, and low-cost to assess detection accuracy across different distances. Results indicated that high-cost devices could accurately detect vehicles up to 500 meters away, whereas medium and low-cost devices exhibited reduced detection accuracy and range. The findings highlight the impact of hardware specifications and environmental conditions on system performance. The proposed approach addresses limitations of conventional traffic monitoring by providing an adaptable, open-source infrastructure that reduces hardware costs while ensuring real-time processing. Utilizing mobile devices enhances scalability and cost-effectiveness compared to traditional RSUs, which are expensive and hard to deploy at scale. Future research will integrate functionalities like pedestrian detection and vehicle tracking to further enhance smart transportation systems. This study demonstrates the feasibility of Android-based RSUs, offering a practical alternative to conventional methods and advancing intelligent traffic management solutions.