A Multi-Agent Prediction Method for Data Sampling and Transmission Reduction in Internet of Things Sensor Networks

Area of Science:

• Computer Science
• Electrical Engineering
• Data Science

Background:

• Sensor networks are crucial for Internet of Things (IoT) applications, providing real-time data.
• Limited network bandwidth, storage, processing, and energy necessitate efficient data handling in IoT sensor networks.

Purpose of the Study:

• To introduce a new method for reducing data transmission in IoT sensor networks.
• To decrease the amount of data samples collected by sensor nodes by predicting sensor reading intervals.

Main Methods:

• A multi-agent system is employed to determine predicted intervals of possible sensor readings.
• Agents independently analyze historical data to evaluate similarities between past and current sensor readings for prediction.
• The prediction algorithm is executed at the IoT gateway or in the cloud.

Main Results:

• The method effectively suppresses unnecessary transmissions and reduces collected data samples.
• Experimental results demonstrate improved accuracy of prediction intervals.
• A higher rate of transmission reduction was achieved compared to existing prediction methods.

Conclusions:

• The proposed method is suitable for IoT sensor networks with resource-constrained, low-end devices.
• It efficiently manages data by determining the usefulness of sensed data and optimizing transmission frequency.
• The approach enhances prediction accuracy and significantly reduces data transmission in IoT sensor networks.