State of Charge Estimation for Both Electric Bus and Passenger Vehicles with Different Battery Types Using Multi-Instance Learning

Real-World SoC Estimation for Lithium-Ion Batteries in Electric Buses: Sampling Resolution and Machine Learning Efficiencys

This paper presents a deployment-oriented evaluation of data-driven SoC estimation using large-scale operational data from a battery electric bus, showing that temporal resolution strongly affects the accuracy-efficiency trade-off.

Hybrid neural network-based state of health estimation for lithium-ion batteries in electric vehicles

Lithium-ion batteries fundamentally influence the reliability, safety, and service life of electric vehicles, making accurate State of Health (SOH) estimation essential for advanced battery management systems. To address the nonlinear, multiscale, and usage-dependent degradation patterns found in real-world cycling, this study proposes a hybrid neural network framework that combines convolutional feature extraction with recurrent temporal modeling. The architecture incorporates mathematically defined convolutional modules for localized voltage–capacity pattern recognition, recurrent units for longterm degradation dynamics, and a fusion mechanism supported by multi-metric loss functions. Using publicly available datasets—including National Aeronautics and Space Administration (NASA) cells B0005/B0006 and the Oxford Battery Degradation Dataset—the method is evaluated under diverse degradation trajectories characterized by mid-life plateaus, nonlinear capacity fade, and accelerated end-of-life decline. Results demonstrate that the hybrid model exhibits improved robustness to measurement noise, partial cycling, and operational variability compared with single-path architectures, thereby providing more stable and adaptable SOH predictions. These findings indicate that hybrid neural network frameworks offer a promising technical pathway for next-generation battery management systems, enabling more reliable SOH monitoring, enhanced predictive maintenance strategies, and longer-lasting electric vehicle battery performance.

A deep learning approach for electric vehicle battery charging duration estimation using IoT

The rapid development of electric vehicles (EVs) has increased the desire for precise charging duration estimate to enhance charging station administration and elevate customer experience. This research presents a deep learning (DL) architecture that uses internet of things (IoT)-enabled data to categorise charging duration into short, medium, and long classifications. The assessment dataset comprises many numerical and categorical variables affecting battery performance and charging behaviour, providing a thorough foundation for prediction. The system utilises a deep neural network (DNN) architecture with nonlinear transformations and regularisation techniques, trained with adaptive optimisation to provide durable convergence. Extensive experiments indicate that the proposed model achieves an overall accuracy of 99.26%, markedly improving traditional machine learning (ML) techniques. These results highlight the potential of DL to adeptly discern complex linkages in charging dynamics, providing dependable predictions for duration classification. The framework provides an advanced basis for implementation in smart charging infrastructures, facilitating effective scheduling, minimizing waiting times, and endorsing predictive maintenance measures. It enhances the reliability and efficiency of EV charging ecosystems with data-driven, IoT-enabled DL technologies.

Co-optimization of DC-bus Control and Battery Health Management for Reliable Solar-powered Electric Vehicle Charging Stations

AI driven real-time battery SOC estimation with intelligent power control

Accurate estimation of battery State of Charge (SOC) is essential for the reliable and efficient operation of Battery Management Systems (BMS) in electric vehicles, renewable energy storage, and portable electronics. Conventional techniques such as open-circuit voltage measurement and Coulomb counting suffer from limitations including sensitivity to operating conditions, cumulative errors, and reduced accuracy under dynamic loads. This paper presents an integrated hardware–software framework for real-time SOC estimation and intelligent power control using a machine learning approach. An Arduino-based sensing layer acquires voltage, current, temperature, and discharge time, which are processed using a hyperparameter-optimized Random Forest regression model implemented in Python. The predicted SOC is transmitted to an ESP32 microcontroller that executes a threshold-based load management strategy using optocoupler-isolated relay switching. Experimental results demonstrate high prediction accuracy with a Mean Absolute Error of 0.0062% and R² of 99.9997%, outperforming conventional and deep learning-based approaches. The proposed system offers a low-cost, scalable, and IoT-enabled solution for precise battery monitoring and adaptive load control in real-time applications.

A Review of State-of-Charge Estimation Methods for Lithium-Ion Batteries Based on Improved Particle Filtering Algorithms

Lithium-ion batteries have become the core energy carrier of electric vehicles and energy storage systems due to their high energy density and long cycle life. Their state of charge (SOC) is an important parameter in battery management systems, playing a key role in energy management, safety protection, and life prediction. However, the SOC cannot be measured directly, and it is difficult for traditional estimation algorithms to balance accuracy and real-time under nonlinear, non-Gaussian noise, multiworking conditions, and parameter time-varying conditions. This paper reviews the research progress of SOC estimation based on an improved particle filter (PF); systematically analyzes its comparison with direct measurement, data-driven, physical model, and mixed methods; and focuses on the fusion path of improved PF and the equivalent circuit model, online parameter identification, intelligent optimization algorithm, and deep learning. The results show that the improved PF algorithm can effectively alleviate the problem of particle degradation and significantly improve the estimation accuracy and robustness under the whole life cycle and complex working conditions. Among them, the nonlinear autoregressive neural network combined with particle filtering method has the best performance, achieving a root mean square error of about 0.02% and a maximum error of less than 0.07% under dynamic working conditions, which is significantly better than other methods. The results show that improving PF can not only break through the bottleneck of traditional methods but also provide an important direction for future high-precision SOC estimation. This paper suggests that subsequent research should further focus on intelligent optimization, multisource fusion, and embedded lightweight implementation to promote the large-scale engineering application of improved PF methods in electric vehicles and energy storage systems.