Indoor Environment Prediction with Multiple Sensors and Generative AI via MCP Integration

Abstract

This paper demonstrates an integration of Large Language Model (LLM), a powerful text-centric Generative AI (GenAI) with an Internet of Things (IoT) sensor network for advanced environmental monitoring and analysis. This paper presents a novel system architecture that integrates a multi-sensor IoT network with a Generative AI model using the Model Context Protocol (MCP) for real-time indoor environmental prediction. MCP is used to orchestrate data flow from multiple sensors such as temperature, humidity, and carbon dioxide to GenAI for analysis and prediction of indoor air quality. The AI-driven insights are then delivered to users through a web application. The evaluation results confirmed the system's high performance, achieving an 85% average prediction accuracy across all three metrics, calculated based on whether predictions fell within predefined tolerance levels. This work establishes the practical value of MCP in a real-world application and showcases the potential of GenAI to transform multi-point sensor data into predictive insights.