DeepEdge: A Novel Appliance Identification Edge Platform for Data Gathering, Capturing and Labeling

Area of Science:

• Electrical Engineering
• Computer Science
• Energy Systems

Background:

• The Internet of Things (IoT) for smart grids necessitates effective appliance monitoring.
• Appliance identification is crucial for monitoring but existing cloud-based platforms are resource-intensive.
• There is a need for cost-effective, edge-based solutions for appliance identification.

Purpose of the Study:

• To propose a novel edge identification platform for appliance monitoring in smart grids.
• To develop a low-cost platform for data gathering, capturing, and labeling of appliance usage.
• To evaluate the performance of the proposed edge platform in terms of accuracy and impact on load disaggregation.

Main Methods:

• Development of a novel edge computing platform for appliance identification.
• Implementation of data gathering, capturing, and labeling functionalities at the edge.
• Experimental validation of the platform's performance and its effect on non-intrusive load disaggregation.

Main Results:

• The proposed edge platform achieved an average appliance identification accuracy of 98.5%.
• The platform demonstrated improved accuracy for non-intrusive load disaggregation algorithms.
• The edge-based approach offers a low-cost alternative to cloud-based solutions.

Conclusions:

• The developed edge platform is effective for appliance identification in smart grid applications.
• This edge solution reduces computational resource requirements compared to cloud-based systems.
• The platform enhances the accuracy of load disaggregation, contributing to smarter energy management.