Secure Estimation Against Malicious Attacks for Lithium-Ion Batteries Under Cloud Environments

This paper is concerned with the secure estimation problem for the state of charge of Lithium-ion batteries subject to malicious attacks during the data transmission from sensors to cloud-based battery management system terminal. First, the second-order resistance-capacitance equivalent circuit model, whose parameters are identified by Kalman filter in an off-line manner, is introduced to describe the internal dynamics of lithium-ion batteries. Then, by applying the chi(2) detection mechanism, real-time malicious attacks are first detected and then a secure estimator is designed to suppress the influence of attacks on the estimation performance. An upper bound of the filtering error covariance is determined by solving certain coupled Riccati-like equations, and the filter parameter is obtained by minimizing such an upper bound at each time step. Finally, the validity of the proposed attack detection approach and the effectiveness of the developed estimation scheme are verified by experiment results under Federal Urban Driving Schedule condition.

Automated summary

This paper addresses the issue of secure estimation for the state of charge of Lithium-ion batteries in the context of malicious attacks during data transmission from sensors to cloud-based battery management system terminal. It introduces a second-order resistance-capacitance equivalent circuit model for describing the internal dynamics of lithium-ion batteries, with parameters identified using Kalman filter in an off-line manner. The paper then proposes a chi(2) detection mechanism to detect real-time malicious attacks, and designs a secure estimator to mitigate the impact of attacks on the estimation performance. Experimental results under Federal Urban Driving Schedule condition validate the effectiveness of the proposed attack detection approach and estimation scheme.