Adaptive Control Model for Indoor Environment of Intelligent Buildings Integrating Multi-Source Sensing Data

Improving energy efficiency and occupant comfort in intelligent buildings is a pressing challenge, given the increasing complexity of indoor environments and the proliferation of multi-source sensor data. This study aims to develop an adaptive control framework that leverages distributed sensing and advanced computational intelligence to address these challenges. The proposed method integrates real-time data from temperature, humidity, CO2, and occupancy sensors, employing normalization and anomaly detection to enhance data quality. A centralized Kalman filter is used for optimal state estimation and noise reduction, while reinforcement learning algorithms drive dynamic optimization of HVAC settings to balance energy consumption and occupant comfort. High-fidelity simulation experiments demonstrate that the framework reduces average energy consumption by 23% compared to traditional PID and fuzzy logic controllers, while maintaining optimal comfort levels for over 90% of occupancy time. Additionally, the system exhibits strong resilience to sensor faults and environmental disturbances, ensuring robust performance in realistic building scenarios. These findings confirm that fusing multi-source sensor data with adaptive learning-based control provides a practical and scalable solution for next-generation smart building environments.