A Battery State of Health Prediction Framework Considering User Behavior for On-road Electric Vehicles

This paper proposes a battery state of health (SOH) prediction framework based on a field dataset of 200 on-road electric vehicles collected during June 2022 to June 2023. 21 user charging and driving behavior features are extracted from the field data. A long short term memory (LSTM) neural network is employed to predict the battery SOH. The impacts of the 21 features on the battery SOH degradation are ranked.