Abstract: Lithium-ion batteries, as one of the core components of new energy vehicles, have become a key research field for new energy vehicles. To pursue more efficient battery management technologies, the estimation method of battery state of charge (SOC) has become a critical research content. At present, mainstream battery SOC estimation methods include data-driven technologies, advanced filtering methods, and machine learning algorithms. Advanced filtering methods and machine learning algorithms exert remarkable effects on improving the accuracy of SOC estimation, and the integration of artificial intelligence and hybrid models also demonstrates favorable performance in enhancing SOC estimation performance. This paper conducts a specific evaluation on battery SOC estimation methods in battery management systems, summarizes and compares existing SOC estimation methods, concludes the advantages and disadvantages of various methods, proposes a system architecture for multi-state joint estimation, and forecasts the development trends of SOC estimation methods, so as to contribute to the safe operation and efficient management of lithium-ion batteries.

Key words: new energy vehicles; lithium-ion batteries; battery management system; SOC estimation

摘要： 锂离子电池作为新能源汽车的核心部件之一，已成为新能源汽车重点研究领域。各企 业为了追求更高效的电池管理技术，电池荷电状态（SOC）估算方法成为关键研究内容。目 前电池 SOC 估算方法主要有数据驱动技术、先进的滤波方法以及机器学习算法等，先进的滤 波方法和机器学习算法在提高 SOC 估算的准确性方面效果显著，人工智能与混合模型的集成 在提高 SOC 估算性能方面也显示出良好效果。文章通过对电池管理系统中的电池 SOC 估算 方法进行具体评估，总结现有的 SOC 估算方法并对这些方法进行比较，归纳出各类方法的优 势及不足，提出了一种多状态联合估算的系统架构，并预测了 SOC 估算方法的发展趋势，以 期为锂离子电池的安全运行和高效管理作出贡献。

关键词: 新能源汽车；锂离子电池；电池管理系统；SOC 估算