SmartBins: AI-Driven Predictive Waste Management for Chennai Municipal Solid Waste
DOI:
https://doi.org/10.46647/rdems0207006Keywords:
Smart waste management, IoT, Random Forest, Isolation Forest, VRPTW, digital twin, Chennai, Municipal Solid Waste, real-time dashboard, anomaly detection.Abstract
SmartBins is an AI-driven, context-aware predictive waste management system designed for the Greater Chennai Corporation (GCC) bin network. The system integrates a high-fidelity digital twin simulator calibrated to official TNPCB and GCC statistics, Random Forest fill-level forecasting, Isolation Forest anomaly detection, priority-weighted Vehicle Routing Problem with Time Windows (VRPTW), crowdsourced citizen validation, and a complete real-time Streamlit dashboard. A 30-day simulation over 120 bins (86,400 hourly readings) achieves a 68% reduction in overflow events (648 vs. 2,027 for the reactive baseline), services 2.5x more at-risk bins per route (64 vs. 26), and attains R² = 0.998 with MAE = 0.0047. All four research hypotheses are empirically supported through quantitative evaluation on real GCC bin locations.