Abstract: In order to accurately estimate the state of health (SOH) of lithium-ion batteries, this paper proposes an SOH evaluation method based on an improved particle filter algorithm. In order to solve the problem that the particle weight tends to zero in the traditional particle filter algorithm and leads to the loss of particle diversity, the residual resampling algorithm is introduced to replace the traditional resampling method by separating the integer and decimal parts of the particle weight, so as to reduce the particle degradation phenomenon and maintain the diversity of the particle set. At the same time, the unscented Kalman filter (UKF) algorithm is combined to generate Sigma points based on state mean and covariance to capture the uncertainty of the system state more accurately and avoid the truncation error of local linearization approximation. The experimental data published by NASA laboratory are used for verification, and the results show that compared with the traditional particle filter algorithm, the proposed method reduces the average error to less than 2%, and significantly improves the accuracy and robustness of SOH estimation.

Key words: lithium battery; residual resampling; unscented Kalman filter; particle filter; state of health

摘要： 为了准确估计锂离子电池健康状态（SOH），文章提出一种基于改进粒子滤波算法的 SOH 评估方法。针对传统粒子滤波算法中粒子权重趋于零、导致粒子多样性丧失的问题，引 入残差重采样算法，通过分离粒子权重的整数和小数部分，以替代传统的重采样方法，从而 减轻粒子退化现象，保持粒子集的多样性。同时，结合无迹卡尔曼滤波（UKF）算法生成基 于状态均值和协方差的 Sigma 点，以更精确地捕捉系统状态的不确定性，避免局部线性化近 似的截断误差。采用 NASA 实验室公布的试验数据进行验证，结果表明，与传统粒子滤波算 法相比，该方法将平均误差降低至 2%以内，显著提升了 SOH 估计的精度和鲁棒性。

关键词: 锂电池；残差重采样；无迹卡尔曼滤波；粒子滤波；健康状态