Answers are grounded in this paper and cite their sources.
Outdated passenger monitoring systems in Philippine public transportation have led to operational inefficiencies, route congestion, and limited commuter decision-making, particularly in modern jeepneys that lack intelligent occupancy tracking. This study addresses these challenges by developing an AI-powered object detection system for real-time passenger counting, aiming to enhance commuter experience, optimize fleet utilization, and support urban mobility planning. The research employed a three-phase developmental design: (1) AI model development, wherein the YOLOv10 object detection algorithm was trained on a curated dataset of 3,000 manually annotated images extracted from real-world CCTV footage in Bacolod City modern jeepneys; (2) mobile application development, featuring a commuter-centric application built with Django and Bootstrap that integrated real-time GPS tracking, occupancy status indicators (Empty, Half-Full, Full), and an interactive map interface; and (3) system evaluation by experts using McCall’s software quality model. The AI model demonstrated strong performance, achieving 94.2% precision, 90.8% recall, and 96.1% mean average precision (mAP), validating its robustness in operational environments. Expert assessments affirmed the system’s excellence in usability, flexibility, and maintainability. The findings highlight the viability of AI-driven occupancy monitoring as a transformative solution for advancing commuter-centric public transportation systems in developing urban contexts.
No takes yet. Share an insight, caveat, or question.
Daba-ay et al. (2025) studied this question.