Total: 1
Accurate and stable state estimation is critical for battery management. Although dual Kalman filtering can jointly estimate states and parameters, the strong coupling between filters may cause divergence under large initialization errors or model mismatch. This paper proposes a Stability Guaranteed Dual Kalman Filtering (SG-DKF) method. A Lyapunov-based analysis yields a sufficient stability condition, leading to an adaptive dead-zone rule that suspends parameter updates when the innovation exceeds a stability bound. Applied to an electrochemical battery model, SG-DKF achieves accuracy comparable to a dual EKF and reduces state of charge RMSE by over 45% under large initial state errors.