A Review of Smart Lithium-Ion Battery Manufacturing: AI-enabled Decision and Control Approaches

With the rapid expansion of the new energy sector and the acceleration of electrification, lithium-ion battery (LIB) manufacturing is undergoing a transition that must jointly prioritize high quality, high efficiency, and sustainability. Conventional practices face challenges in multi-stage coupling and high-dimensional optimization, underscoring the need for intelligent solutions. This review surveys AI(Artificial Intelligence)-enabled decision-making and control in smart LIB manufacturing. After outlining the canonical process flow and the decision–control hierarchy, we distill core requirements and key technical issues. Advances are synthesized across five strands—machine learning, deep learning, reinforcement learning, knowledge-driven approaches, and hybrid AI—covering their theoretical basis, capabilities, and representative deployments in coating, calendering, assembly, and formation. The literature indicates that AI delivers tangible gains in process parameter optimization, quality prediction, defect inspection, and production scheduling, thereby enabling cross-layer intelligent decision-making and real-time control. We further propose a unified analysis framework that maps method families to manufacturing scenarios, providing a panoramic view and side-by-side comparison to support method selection and system design for industrial upgrades. Persistent challenges remain in data quality and governance, model generalization and robustness, and compute and latency constraints for edge deployment. Looking ahead, promising directions include digital-twin-driven co-simulation and control, knowledge-augmented reasoning, and system-level integration of hybrid AI. This review aims to inform the systematic deployment of AI across LIB production and to guide both theoretical research and engineering practice in smart manufacturing.