Hybrid State-of-Charge Estimation for Primary Battery Powered Sensor Nodes

Accurate battery state-of-charge (SOC) estimation is crucial to maximize the lifetime of wireless sensor networks (WSNs). Meanwhile, considering the limited capability of sensor nodes, lightweight estimation methods are preferred. Unfortunately, the existing lightweight methods are not satisfactory when facing rapidly changed loads. Moreover, their accuracy is not stable throughout the whole lifetime. This article analyzes the error variation of the existing methods theoretically and finds that these methods are actually complementary in different stages of the discharge process. Therefore, integrating different methods based on the characteristics of target batteries would be beneficial. With this finding in mind, a hybrid estimation method is proposed to realize more accurate and stable SOC estimation, which combines the advantages of current consumption-based methods in the initial stage of discharge process with the advantages of introducing battery terminal voltage in the later stage of discharge process. In order to evaluate the proposed method comprehensively, different primary batteries are employed and different working conditions (constant current, constant resistance, and emulated duty cycle loads) are designed. Experimental results show that the proposed method is more accurate for different batteries and working conditions. In particular, under duty cycle loads commonly used in sensor networks, the estimation error reduces by up to 72.39% and 59.40% for carbon zinc battery and alkaline battery, respectively. Meanwhile, temperature changes will not affect the advantages of the proposed method. More importantly, its estimation accuracy is more stable throughout the whole lifetime.