State of health estimation with an improved Gaussian process regression

State of health estimation is crucial for ensuring the safety and reliability of lithium-ion batteries. However, due to various user's habits, the discharging profiles of the lithium-ion battery are totally different in practice, thereby limiting applications of many feature-based methods in real operations. Compared with the discharging process, the charging process is usually executed in a more peaceable and predictable manner. Therefore, two time-related aging features are extracted from the constant current-constant voltage charging process in this work. By introducing the quantum computing theory into the classical intelligent learning model, an improved Gaussian process regression framework, as well as its application to describe relations between extracted features and the battery SOH, is proposed and illustrated in detail. With datasets of lithium-ion batteries by NASA, experiment and comparison results validate the effectiveness, accuracy and superiority of the presented online SOH estimation framework.