Implementation of Smart Farm Systems Based on Fog Computing in Artificial Intelligence of Things Environments

Area of Science:

• Computer Science
• Agricultural Technology
• Artificial Intelligence

Background:

• Cloud computing faces challenges with massive data transmission in Internet of Things (IoT) applications.
• Fog computing offers a decentralized solution to mitigate cloud computing overhead.
• Smart farming requires efficient data processing for real-time decision-making.

Purpose of the Study:

• To implement and evaluate an Artificial Intelligence of Things (AIoT) system utilizing fog computing in a smart farm environment.
• To assess the impact of fog computing on network traffic volume and communication protocol performance.
• To develop and validate an AI-driven algorithm for determining coffee tree maturity levels.

Main Methods:

• A fog computing-based AIoT system was deployed on a coffee tree farm.
• Network traffic was compared between fog-enabled and non-fog systems.
• Performance of Hypertext Transport Protocol (HTTP), Message Queuing Telemetry Transport (MQTT), and Constrained Application Protocol (CoAP) was evaluated.
• A convolutional neural network (CNN) model was used for coffee fruit maturity classification.

Main Results:

• The fog computing system achieved a 26% reduction in cumulative data volume compared to a non-fog system.
• Message Queuing Telemetry Transport (MQTT) demonstrated stable performance regarding data volume and loss rate.
• The CNN-based algorithm provided reliable results for determining coffee fruit maturity levels.

Conclusions:

• Fog computing effectively reduces data transmission overhead in smart farm AIoT applications.
• MQTT is a suitable protocol for stable data transmission in this context.
• AI and fog computing integration offers a promising approach for smart agriculture optimization.