Research on SOC estimation method of multi-rotor UAV battery based on improved extended Kalman filter

Multi-rotor UAV is a special unmanned rotorcraft with three or more rotor shafts, used in many fields. The Multi-rotor UAV generally uses a battery as a power source, so accurately estimating the Battery State of Charge (SOC) in a Multi-rotor UAV energy management system is crucial, which confirms the Multi-rotor UAV can be operated efficiently and safely. But, because the working condition of Multi-rotor UAV is special, the error of battery SOC is estimated by the current integration method, which has cumulative error. Although the SOC value estimated by the current integration method is verified in actual industrial applications, it does not fundamentally solve the problem that the current integration method relies on the initial value and the accumulated error is large. Kalman filter method estimates battery SOC independent of initial value and does not produce error, and the SOC estimation accuracy can be improved by using the EKF method. Aiming at the special working conditions of the Multi-rotor UAV, the filtering gain of the EKF method is improved for the dynamic adjustment of the filter gain. Effectively improve the tracking performance of the EKF method. The experimental results show that the improved Kalman filter can quickly track the true SOC value when the current mutates, and the tracking effect increases by about 70%.The tracking effect of the estimation process is improved by using the filter gain of the dynamic correction, which improve the real-time performance of the estimated SOC value to ensure that the accuracy of the Multi-rotor UAV SOC value is within 5%, effectively solving the problem of inaccurate estimation of the SOC value of the Multi-rotor UAV in the project.