ARCHIVES

Food spoilage is a major challenge in the food supply chain, resulting in significant economic losses, food waste, and potential health risks. Conventional methods of assessing food freshness rely on manual inspection, which is subjective, time-consuming, and unsuitable for continuous monitoring. This paper presents an AI-based Multi-Parameter Food Spoilage Prediction System that integrates Internet of Things (IoT) technology with machine learning for real-time food quality assessment. The proposed system employs multiple sensors, including an MQ135 gas sensor, DS18B20 temperature sensor, DHT22 humidity sensor, soil moisture sensor, and load cell with an HX711 amplifier, to monitor critical environmental and physical parameters associated with food spoilage. An ESP32 microcontroller acquires and transmits sensor data, while a Random Forest classifier analyzes the fused sensor inputs to classify food into four freshness levels: Fresh, Early Spoilage, Unsafe, and Spoiled, with an estimation of the remaining shelf life. The system also performs cold-chain monitoring by continuously tracking storage temperature and generating alerts when unsafe conditions are detected. Experimental evaluation demonstrates that the integration of multi-sensor data and machine learning improves prediction reliability compared with single-parameter approaches, enabling early spoilage detection and real-time monitoring. The proposed solution offers a scalable, cost-effective, and intelligent framework for enhancing food safety, minimizing wastage, and supporting efficient storage and transportation management in modern food supply chains.