State of charge (SOC) estimation is the core algorithm of the battery management system. However, the commonly used model-based, data-driven, or experiment-based methods struggle to independently achieve accurate SOC estimation under different working conditions and temperatures, which affects battery performance and safety. To this end, this paper proposes an online SOC estimation method that combines the model-driven and double-data-driven approaches. The unscented Kalman filter (UKF) based on the first-order RC model is used to achieve robust SOC estimation, while the data-driven long short-term memory (LSTM) neural network is used to achieve fast SOC estimation. The former model has an excellent dynamic performance and the latter has high steady-state accuracy. The SOC estimation results are input into the SOC estimation model of series LSTM so that the stable but inaccurate SOC values estimated by UKF in the first part and the accurate but fluctuating SOC values estimated by LSTM can be correlated and corrected, achieving a fast and accurate SOC estimation under various working conditions. The estimation results show that the above method has strong robustness and high accuracy, and effectively reduces model complexity and data redundancy. In addition, the root mean square error of SOC estimation under different working conditions is controlled within 1–2.3% at 0 °C, 25 °C, and 45 °C, which is better than the traditional single-SOC estimation method.

The cloud-integrated battery management system (CIBMS) of electric vehicles (EVs) provides an excellent solution for reducing the cost, enabling refined battery management, and implementing advanced algorithms. However, limited network communication bandwidth, non-Gaussian noises, and incomplete measurement information pose significant challenges for CIBMS in accurately estimating the battery states. To overcome these obstacles, this article proposes an event-triggered (ET) robust state of charge (SOC) estimation strategy for CIBMS. First, an ET communication mechanism is designed for CIBMS to ensure that data are transmitted to the cloud on demand. Then, a second-order fractional-order model is introduced, and the discrete-time form of the dynamic equation is derived. Subsequently, the improved fractional-order extended Kalman filter (IFOEKF) is proposed, which reduces the linearization error of measurement equation in FOEKF. Furthermore, for non-Gaussian noises, the robust IFOEKF (RIFOEKF) is developed based on Masreliez–Martin method and adaptive algorithm. Finally, to achieve accurate SOC estimation under incomplete measurement information, the compensation protocol is established and the corresponding ET-RIFOEKF is devised. The proposed method is validated under different EV driving conditions. The experimental results show that the proposed method not only effectively reduces the frequency of transmitting measurement data to the cloud but also has high accuracy and robustness for SOC estimation.

Accurate estimation of the state of charge (SOC) and state of health (SOH) is critical for ensuring the safety and reliability of lithium‐ion batteries, especially under varying temperature conditions. Temperature significantly impacts battery performance, introducing non‐linear effects that complicate state estimation. This paper proposes a novel multi‐scale co‐estimation strategy that integrates a temperature‐dependent fractional‐order model (FOM) with a dual fractional‐order adaptive unscented Kalman filter (DFOAUKF). The FOM is developed based on electrochemical impedance spectroscopy and the temperature‐dependent open‐circuit voltage, capturing intrinsic electrochemical characteristics of the battery. Parameters are identified using a combined offline and online method. The DFOAUKF algorithm leverages the time‐scale differences between SOC and SOH, enabling micro‐time SOC estimation in seconds and macro‐time SOH estimation over test cycles, while adaptively updating noise covariance matrices. Experimental validation under dynamic stress test (DST) and urban dynamometer driving schedule conditions at 10, 25 and 40 demonstrates the robustness and accuracy of the proposed method. Results show that the root mean square errors of terminal voltage remains within 0.039 V, while SOC and SOH estimation errors are less than 1.25% and 0.64%, respectively, across all temperature levels and aging degrees. The findings highlight the effectiveness of the proposed co‐estimation strategy in improving battery state monitoring and lifespan.