Enhanced Unscented Kalman Filter for Accurate State of Charge Estimation in Aerial Drone Lithium-Ion Batteries

Accurate state of charge (SOC) estimation is a fundamental aspect of lithium-ion batteries (LIBs) used in aerial drones. This estimation significantly affects both the control accuracy and the reliability of the energy storage system embedded in the LIB power supply. The Unscented Kalman filter (UKF) algorithm is used to estimate the SOC in LIBs. However, the algorithm's ability to accurately monitor rapid fluctuations in pulse current is compromised, resulting in suboptimal tracking performance and non-negligible estimation errors. This paper presents an improved UKF to address this challenge. The results show that a two-order PNGV (Partnership for a New Generation of Vehicles) model effectively limits the maximum estimation error to less than 0.2 V. Even under significant voltage and current variations, the UKF estimation error can reach up to 2%. However, the use of the enhanced UKF algorithm maintains remarkable stability with only a 1% estimation error. This represents a 0.8% improvement in estimation accuracy compared to conventional methods. These results provide a robust theoretical basis for improving energy management in LIBs for aerial drones.