Lithium-ion batteries SoC estimation using an ANFIS-based adaptive sliding mode observer for electric vehicle applications infrastructures

State of charge (SoC) estimation is a key function in battery management systems (BMSs) that is not directly measurable and should be estimated using estimation methods. Estimating the SoC requires addressing model uncertainty while determining battery model parameters. Robust battery SoC estimation approaches overcome this challenge. Sliding mode parameter estimation chatters in its original form. To solve this problem, this paper adapts the sliding gain switching estimator by an adaptive fuzzy system to solve the chattering problem. A neural network is used to optimise fuzzy systems, which demand optimisation strategies. The research proposes an adaptive neuro-fuzzy SMO for SoC estimation to improve robustness, accuracy, and response chattering. SoC estimation uses a lithium-ion battery cell equivalent circuit model (ECM). The open circuit voltage's nonlinear relationship with charge makes this model nonlinear. The recommended methodology has been tested using a set of software-in-the-loop experiments, which show that chattering has been abolished and accuracy can be decreased by 5% compared to the standard SMO.