A Hybrid Algorithm for High‐Accuracy State‐of‐Charge Estimation in Lithium‐Ion Batteries

Abstract

Accurate state‐of‐charge (SOC) estimation is essential for ensuring the safety and efficiency of lithium‐ion battery systems under complex operating conditions. To address limitations in convergence speed and estimation accuracy, this paper proposes an improved Latin Hypercube Tuna Swarm Optimization (LHTSO) algorithm. The method enhances population initialization via Latin hypercube sampling, incorporates a stage‐adaptive search strategy, and introduces an elite‐guidance mechanism to improve global optimization performance. An integrated LHTSO‐BP‐UKF framework is further developed for SOC estimation. Experimental validation is conducted under multiple driving cycles (Dynamic Stress Test (DST), New European Driving Cycle (NEDC), Federal Test Procedure (FTP), Urban Dynamometer Driving Schedule (UDDS)) and a wide temperature range (−10 to 40 °C). Results demonstrate that the proposed method consistently outperforms conventional Unscented Kalman Filter (UKF) and its variants. Under the challenging DST condition at 25 °C, the proportion of samples with estimation error exceeding 1% is reduced from 80.20% to 4.10%, achieving a 96.55% relative improvement. Moreover, the method maintains stable and bounded estimation under low‐temperature conditions. These results confirm the robustness, generalization capability, and practical applicability of the proposed approach.