Battery state of charge estimation for electric vehicle based on hybrid barnacles mating optimizer-feed forward neural network

Z Mustaffa and MH Sulaiman, JOURNAL OF THE CHINESE INSTITUTE OF ENGINEERS, 48, 1196-1216 (2025).

DOI: 10.1080/02533839.2025.2517349

Accurately determining battery charge levels remains a significant challenge in electric mobility, particularly given the complex electrochemical processes and varying operational conditions that affect battery behavior. This research introduces an innovative approach combining Feed Forward Neural Networks (FFNN) with the Barnacles Mating Optimizer (BMO) technique to enhance battery charge level predictions, addressing the limitations of conventional estimation methods through advanced computational intelligence. The key innovation lies in leveraging BMO's capabilities to fine-tune the FFNN's parameters, resulting in improved of State of Charge (SoC) estimation precision. The integrated system was evaluated using operational data collected from a BMW i3 across 70 different journeys. Performance assessment utilized three distinct error metrics: Normalized Mean Square Error (NMSE), Root Mean Square Percentage Error (RMSPE), and Mean Absolute Percentage Error (MAPE). The experimental results revealed superior performance of our BMO-FFNN, achieving error rates of 0.0954 (NMSE), 5.0954% (RMSPE), and 3.7919% (MAPE). These figures demonstrated marked improvement over comparable hybrid approaches incorporating alternative optimization methods such as Salp-Swarm Algorithm (SSA), Moth-Flame Optimization (MFO), and Whale Optimization Algorithm (WOA) when paired with the same FFNN architecture. The demonstrated accuracy of this novel BMO-FFNN suggests promising applications in advancing electric vehicle energy management systems.

Return to Publications page