Energy consumption of Wireless Sensor Networks (WSN) is an important aspect in
the design requirement. This is especially true in a situation where WSN is being
operated in isolated areas and thus relying on batteries due to unavailability of power
infrastructure. Since energy efficiency is the main concern in the deployment of WSN,
the sensor node must keep track of the charge that is left in the battery, commonly
referred as the State of Charge (SoC). To prevent the discontinuation of the operation
of the sensor node from power cut off, it is important to find an analytic model for
the battery’s state of charge. In this paper, an optimized structure of Multi-Layer
Perceptron (MLP) is utilized to obtain a model of the battery state-of-charge in
wireless sensor nodes. Results show the suitability of the method that produces
accurate and simple models, capable of being implemented even in low cost and very
constrained real motes.