A Multi-Paradigm AI Framework for Explainable and Proactive Battery State-of-Health Estimation in Electric Vehicles
Abstract
The State of Health (SOH) estimation of the lithium-ion batteries is essential for safe and efficient electric vehicle operation. Existing techniques often used black-box models, needing interpretability and the interaction of the user. This paper represents a multi-paradigm AI framework that combines deep learning, neuro-symbolic reasoning, conversational AI for explainable and interactive SOH estimation, and hybrid intelligence. A Conv1D model predicts the SOH from the real-world EV Battery charging data, while a domain-specific rules specify the transparent explanations. An uncertainty-aware method identifies uncertain predictions and allows human-in-the-loop improvement. The agentic monitoring element produces proactive alerts, and a conversational interface permits a natural language query. The experimental results show that the framework reaches a Mean Absolute Error of $\mathbf{3. 8 9 9 \%}$, while enhancing interpretability, usability, and reliability. The results show that the action of combining multiple AI paradigms can move battery health management beyond standalone predictions toward practical, explainable, interpretable and decision-support systems.