Artificial Neural Networks, Gradient Boosting and Support Vector Machines for electric vehicle battery state estimation: A review

In recent years, Artificial Intelligence has been widely used for determining the current state of Li-ion batteries used for Electric Vehicle applications. It is crucial to have an accurate battery state estimation to prevent over-charging/over-discharging of Li-ion batteries, which would contribute to increased lifetime, thus reducing the running costs of Electric Vehicles. This review paper focuses on some of the commonly proposed Artificial In-telligence data-driven based State of Charge and State of Health estimation that has not been covered in much detail previously. The recent works indexed under Web of Science that used Support Vector Machines, Gradient Boosting and Artificial Neural Networks (with a deep insight on the use of recurrent architectures) for battery state estimation are reviewed. A handful of recent works that implemented their proposed battery state esti-mation algorithm on a hardware prototype is also discussed, along with the current challenges faced in imple-mentation. Since various input features have been suggested for State of Charge and State of Health estimation in the recent literature, a detailed analysis is presented in this paper. Key observations with research gaps are made from the reviewed literature, with identification of major challenges. Future research paths are deduced, with the goal of increasing the feasibility of implementing Artificial Intelligence-based battery state estimation in Electric Vehicles.