IoT Off-Grid, Data Collection from a Machine Learning Classification Using UAV

Area of Science:

• Computer Science
• Electrical Engineering
• Robotics

Background:

• The Internet of Things (IoT) typically relies on commercial electricity and internet infrastructure.
• Off-grid environments present unique challenges for deploying and managing IoT systems due to lack of power and connectivity.

Purpose of the Study:

• To propose and develop an IoT off-grid system for managing utilities and collecting data in environments lacking commercial power and internet.
• To identify the state-of-the-art in off-grid IoT through systematic literature mapping.

Main Methods:

• Utilizing machine learning algorithms for intelligent data selection and filtering.
• Employing a drone for safe and efficient data collection across remote off-grid stations.
• Developing a software architecture for both the drone and off-grid stations.

Main Results:

• A systematic literature mapping identified current advancements in off-grid IoT.
• A software architecture proposal was developed, detailing configurations for drone and station data collection.
• The performance of various machine learning selection algorithms was evaluated on a prototype system.

Conclusions:

• The proposed IoT off-grid architecture effectively addresses data collection and utility management challenges in remote areas.
• Machine learning and drone technology are crucial for optimizing data handling in resource-constrained environments.
• The developed prototype demonstrates the feasibility of the proposed system.