AI-driven and sensor-based composting in a connected bioreactor: A novel framework for predicting maturity and quality dynamics

The intensification of agricultural and urban activities has led to a substantial rise in organic waste generation, underscoring the urgent need for sustainable treatment strategies. This study presents an integrated framework that combines a low-cost, sensor-equipped composting bioreactor with advanced machine learning (ML) models to predict compost maturity and quality in real time. Organic substrates from weekly stock markets and university canteen food waste were co-composted with manure and sawdust under controlled conditions in a smart bioreactor. An embedded IoT-based monitoring system to monitor temperature, moisture, CO2, NH3 emissions enables continuous data acquisition and cloud-based visualization. Simultaneously, a curated dataset from peer-reviewed composting experiments was fused with the real-time sensor data to train supervised ML models, including XGBoost, Support Vector Machine (SVM), Random Forest (RF) and, for both classification of compost maturity stages and regression of the germination index (GI).