AI-Powered Non-Intrusive Load Monitoring (NILM)

This project focuses on Non-Intrusive Load Monitoring (NILM), a technique that identifies the energy consumption of individual household appliances using only a single aggregate power meter, rather than installing separate sensors on each device. Unlike Intrusive Load Monitoring (ILM), which requires per-appliance metering hardware, NILM is more cost-effective, easier to deploy, and better suited to existing smart-meter infrastructure, making it a practical solution for large-scale residential and commercial energy monitoring. By disaggregating total power signals into appliance-level insights, NILM supports demand-side management, energy efficiency programs, and consumer feedback systems that encourage reduced electricity usage.

This research aims to develop AI-based NILM models specifically optimized for deployment on resource-constrained edge devices such as smart meters and microcontrollers. The focus is on building lightweight, computationally efficient architectures that maintain high disaggregation accuracy while minimizing memory footprint and processing overhead, enabling real-time, on-device energy monitoring without dependence on cloud computing or high-frequency sampling infrastructure.