This project investigates energy consumption patterns using AI-based automation compared to standard Automation in a household setup. The focus is on optimizing energy usage for appliances such as lights and fans.
The study primarily involved physical setups using the Atmega328p microcontroller to gather real-world data over two months. Simulations were conducted to validate hardware configurations before deployment.
The full findings and methodologies are documented in a published thesis. For detailed information, please refer to the complete thesis here.
Key aspects of the data collection process include:
-
Physical setups using the Atmega328p microcontroller to control lighting and fan systems.
-
Real-world data capturing energy usage across various phases: pre-override, manual override, and post-override.
-
Energy consumption was calculated using the formula:
Energy (kWh) = Voltage (V) × Current (A) × Time (hours) -
Simulations were used to test hardware scenarios before actual implementation.
All datasets, including hourly and daily energy consumption, are stored in CSV files within the /data directory.
Key findings from the project:
- The AI automation system achieved noticeable energy savings compared to standard Automation, particularly during the manual override and post-override phases.
- AI automation adapted to human behavior patterns to optimize energy usage while maintaining comfort.
- The
/resultsfolder contains graphs and tables showcasing energy consumption trends, comparisons, and cumulative savings.