Context-aware adaptive Send-on-Delta for traffic saving in sensor networks

Abstract

Wireless Sensor Networks (WSNs) are widely deployed for real-time monitoring in domains such as industrial automation, smart buildings, and the Internet of Things (IoT). However, continuous sensor data transmission often results in excessive network traffic and elevated energy consumption, particularly in resource-constrained environments.

This paper presents a context-aware adaptive data reduction strategy based on the Send-on-Delta (SoD) transmission scheme, enhanced with a dynamic threshold adjustment mechanism driven by Bollinger Bands. A key novelty of the proposed approach lies in its global adaptation strategy: instead of configuring transmission thresholds independently at each node, the system evaluates the collective behavior of the network to dynamically reconfigure the local thresholds in a coordinated manner.

The method is supported by a formal and algorithmic model that captures the network-wide adaptation process, and is validated through a comprehensive evaluation including: a real-world deployment with ESP32-based wireless sensors operating over a 24-hour period; a scalability study with up to 40 simulated nodes derived from real data traces; and a post-hoc comparison against representative SoD variants, such as Send-on-Area and Predictive SoD. Experimental results show that the proposed strategy achieves over 89% reduction in transmission volume while maintaining bounded error, outperforming traditional approaches in lossless, context-sensitive scenarios. These results confirm the method’s effectiveness, robustness, and scalability for energy-efficient communication in WSNs.