**4. Conclusions**

This research paper has opened a new Li-ion battery modeling research direction in the HEV BMS applications field by performing several investigations on ARX and ANFIS alternative accurate battery models with a high impact on improving the battery SOC estimators' accuracy and their robustness, design, and real-time implementation in MATLAB and Simulink environments.

The effectiveness of the modeling and SOC estimation strategies is demonstrated through an extensive number of simulations in a MATLAB R2021b software environment. The preliminary simulation results are encouraging, and extensive investigations will be done in future work to extend the applications area. The performance analysis from the last section reveals that ANFIS battery models overpass the second-order linear ARX polynomial battery model in terms of SOC and terminal voltage accuracy and by their capability and suitability to simplify the battery model structure and build robust and accurate SOC Li-ion battery estimators with a high terminal voltage prediction accuracy. The AEKF SOC estimator accuracy based on combined ANFIS model structure is also very accurate compared with AEKF SOC estimator based on ARX dynamic part model with the SOC absolute value lower than 1%, better than the usual 2% SOC value reported in the literature field. Both alternative models are based only on the measurement input-output data set collected by a data acquisition (DAQ) system incorporated in the BMS of HEVs. Besides, the battery SOC and output voltage signals' accuracy is not affected by noise as long as the AEKF SOC estimator is very robust.

*Investigations of Using an Intelligent ANFIS Modeling Approach for a Li-Ion Battery… DOI: http://dx.doi.org/10.5772/intechopen.105529*
