Title: Artificial Intelligence Across Disciplines: Research, Innovation, and Intelligent Solutions
Editors: Dr. Subita Bhagat, Dr. A. Balamurugan, Dr. P. Krishna Kumar, and Mrs. S. Nandhini Devi
ISBN: 978-81-69857-83-3
Chapter: 2
DOI: https://doi.org/10.59646/815/02
Author: H. Pushpakalavalli
Abstract
Autonomous smart environments integrate pervasive Internet of Things (IoT) sensing topologies with advanced artificial intelligence control loops to regulate multi-zone indoor microclimates while curtailing electrical utility burdens. Legacy building management automation relies predominantly on static rule-based scheduling and decoupled proportional-integral-derivative loops that cannot adapt to non-linear thermodynamic interactions, varying occupant densities, and volatile pricing dynamics. This chapter introduces an end-to-end edge-to-cloud automation framework that couples continuous multi-modal environmental telemetry with a continuous-action Proximal Policy Optimization (PPO) reinforcement learning engine. Using an empirical multi-zone commercial facility as a representative testbed, the system monitors spatial dry-bulb temperature, relative humidity, carbon dioxide concentrations, and localized occupant counts to modulate variable-air-volume terminals and central air handling units dynamically. Across a comprehensive 30-day continuous evaluation period, the proposed paradigm achieves a 24.8% reduction in aggregate HVAC energy dissipation, lowers Predicted Mean Vote thermal comfort violations by 78.3%, and prevents indoor air quality non-compliance, demonstrating operational viability for modern autonomous infrastructure.
Keywords: Smart Environments, Internet of Things, Deep Reinforcement Learning, Proximal Policy Optimization, Thermal Comfort, Energy Optimization.