Smart Coulomb Counter for Self-Metering Wireless Sensor Nodes Consumption

Accurate Online Energy Consumption Estimation of IoT Devices Using Energest

“…Power consumption of a sensor node depends on many factors, such as the states of the radio, microcontroller (MCU), Light-emitting diodes (LEDs) and other components [10]. Different mechanisms have been used before to estimate energy consumption on small IoT devices such as Contiki-NG's built-in funcionality, named Energest 2 .…”

“…Different mechanisms have been used before to estimate energy consumption on small IoT devices such as Contiki-NG's built-in funcionality, named Energest 2 . It is a time-based estimation mechanism implemented as a collection of functions and macros that runs directly on the sensor nodes and measures the accumulated time the sensor node spends in different MCU and radio states [10] [9].…”

“…Time-based energy estimation is very easy to implement and configure with existing applications, but the accuracy of the obtained results is not always good enough. Possible problems about the accuracy of the results are shown by Steinfeld et al [10]. For example, the CPU mode is activated whenever the node is active, including the period when the radio transmits or receives some data, but not for all the time needed for a transmission.…”

“…It is very hard to know the accurate period of the CPU and the LPM state of the MCU during one transmission and almost impossible to estimate the energy consumption of the MCU and the radio separately. Steinfeld et al [10] propose a hardware-based approach, named Smart Coulomb Counter (SCC), that easily adds to a sensor node the capability of measuring its energy consumption. Also, they compare their solution with Energest and get very similar results between SCC and Energest in the OFF state, but in the ON state, the consumption values differ greatly (more than 85% of relative error) because of the LED current consumption.…”