A Low-power Embedded IoT System for Accurate Detection of Unauthorized Manhole Access

The recent widespread theft of telecommunication copper cables has significantly threatened the underground network infrastructure in Malaysia. In response, various smart manhole monitoring systems have been developed. This study focuses on implementing a low-power, high-accuracy smart manhole cover monitoring system to extend battery life and enhance theft detection. The proposed system employs the STM32L4 microcontroller, utilizing deep-sleep modes to drastically reduce power usage by deactivating unused peripherals during inactivity, significantly boosting power efficiency given the predominant dormancy of the manhole cover. The manhole cover activity is monitored using a single LSM6DSO Inertial Measurement Unit (IMU) with pre-and post-filtering techniques to improve the response characteristics. Additionally, the system integrates an IoT-based alert mechanism through an SX1278 LoRa transceiver, reporting manhole status to a gateway, and uses a Texas Instrument BQ25570 module for efficient power management, delivering a consistent 2.574V from a single 3.7V Li-Ion battery. In the conducted experiments, the IMU sensor exhibited robust precision with a maximum tilt angle error of $\pm \mathbf{3}^{\circ}$ when subjected to intense shaking, ensuring reliable accuracy under typical ambient conditions characterized by low-to-moderate vibrations. The alert mechanism also proved effective in reporting the manhole cover status, even in enclosed spaces without a direct line of sight to the gateway. Moreover, the system showed a low average current consumption of $\mathbf{19.1}\boldsymbol{\mu} \mathbf{A}$ when operated at a buck-regulated voltage of 2.574V from a 3.7V 3350mAh battery, leading to an estimated theoretical lifespan of up to 23 years for a conversion efficiency of 90%.