Method of optimization of energy supply of iot system for monitoring climatic indicators

Abstract

The article addresses the urgent scientific and applied problem of enhancing the energy efficiency of autonomous climate monitoring systems based on Internet of Things (IoT) technologies. The aim of the study is to develop a comprehensive method for adaptive control of power consumption modes, which minimizes the depletion of the power source without compromising data informativeness. It is substantiated that traditional telemetry collection methods using a fixed schedule are suboptimal for dynamic environments, as they result in excessive radio path activity during periods of parameter stability. To achieve this goal, an energy consumption model for the node is proposed and mathematically grounded, taking into account the dynamics of the microcontroller's transient processes. An algorithm for dynamically adjusting sleep intervals is developed based on calculating the gradient of the measured value: the polling frequency automatically decreases if environmental parameters remain stable. The practical implementation of the method is realized as a universal hardware-software complex based on the ESP32-WROVER-E SoC. A key feature of the architecture is the use of an energy-efficient ULP (Ultra Low Power) coprocessor for background monitoring of the BME680 sensor via a software I2C interface, which allows offloading the main computing core. The software is implemented based on the FreeRTOS real-time operating system using the Finite State Machine (FSM) pattern, which manages transitions between deep sleep, data collection, and active communication states. Experimental research results confirmed the effectiveness of the proposed approach. It was established that the application of the adaptive algorithm reduces the average current consumption of the system in active mode to 12.47 mA (using the CoAP protocol) and to 12.63 mA (for LoRaWAN). This ensures a reduction in total energy consumption by 72–87% compared to baseline fixed-schedule algorithms. It is proven that the proposed architecture significantly extends the autonomous operation time of monitoring devices, ensuring a balance between data detail and energy conservation.