State of Charge Estimation for Power Battery Base on Improved Particle Filter

In this paper, an improved particle filter (Improved Particle Swarm Optimized Filter, IPSO-PF) algorithm is proposed to estimate the state of charge (SOC) lithium-ion batteries. It solves problem inaccurate posterior estimation due degradation. The divides population into three parts and designs different updating methods realize self-variation mutual learning particles, which effectively promotes global development avoids falling local optimum. Firstly, a second-order RC equivalent circuit model established. Secondly, parameters are identified by swarm optimization algorithm. Finally, verified under four driving conditions. results show that root mean square error (RMSE) within 0.4% conditions, maximum (ME) less than 1%, showing good generalization. Compared with EKF, PF, PSO-PF algorithms, IPSO-PF significantly improves accuracy SOC, verifies superiority