Plant factories with artificial lighting (PFALs) are increasingly viewed as a promising solution for resource-efficient and climate-resilient food production. As PFAL systems become more complex, their sustainable development depends not only on advances in individual technologies, but also on the integration of sensing, modeling, and control into closed-loop intelligent systems. This review synthesizes recent progress in intelligent PFAL research from three perspectives: multi-modal sensing, predictive modeling, and control. The literature shows a clear transition from single-factor monitoring to phenotype-oriented perception, from empirical and purely data-driven prediction to hybrid and digital-twin-based modeling, and from static optimization to adaptive and learning-based control. However, progress remains limited by fragmented datasets, weak standardization, limited transferability, insufficient deployment-oriented validation, and the persistent Sim-to-Real gap. In response to these challenges, this review proposes a socio-technical framework for intelligent PFALs that connects technical development with environmental, economic, and policy constraints. Based on this synthesis, four priorities are identified for future research: standardized benchmarks, decision-grade predictive models, safe and transferable intelligent control, and sustainability-oriented system integration. Together, these directions provide a foundation for advancing PFALs toward scalable and sustainable operation.
Zhao et al. (2026) studied this question.