From predictive AI to autonomous food bioprocessing: a critical review of real-time quality monitoring, energy optimization, digital twins, and intelligent process control

Food bioprocessing is transitioning from conventional automation towards intelligent, data-driven, and increasingly autonomous manufacturing. A substantial gap nevertheless remains between accurate process prediction and reliable autonomous control. This critical narrative review examines the convergence of real-time sensing, artificial intelligence (AI), energy and process optimisation, hybrid physics–AI modelling, digital twins, and intelligent closed-loop control in food bioprocess engineering, with explicit attention to what has been demonstrated experimentally, at pilot scale, or only in simulation versus what remains conceptual. A structured narrative review informed by SANRA principles was conducted using Scopus, Web of Science, ScienceDirect, IEEE Xplore, and Google Scholar, with literature published up to 17 August 2026 considered and supplemented by a second, targeted search cycle addressing process analytical technology, electronic sensing, convolutional-network quality evaluation, and digital-twin applications in thermal, baking, and cold-chain processes. Evidence was synthesised according to the technological progression from monitoring and prediction to optimisation, digital twins, intelligent control, and autonomy. The evidence indicates substantial advances in smart sensing, machine learning, multimodal quality assessment, and multi-objective optimisation, with demonstrated opportunities to improve product quality, energy efficiency, productivity, and resource utilisation. Nevertheless, most applications remain concentrated at monitoring, predictive-modelling, and offline-optimisation stages, while validated industrial closed-loop and self-learning systems remain comparatively limited, and quantitative sustainability and economic evidence specific to autonomous food bioprocessing remains scarce. Digital twins and hybrid physics–AI approaches provide an important bridge towards autonomy by integrating physical process knowledge, real-time data, prediction, optimisation, and feedback control. This review proposes a six-level Food Bioprocess Autonomy Framework (FBAF), progressing from manual processing to self-learning autonomous bioprocessing, and situates it explicitly against ISA-95, RAMI 4.0, and SAE J3016-derived autonomy scales, providing measurable per-level criteria and illustrative classification of published systems. Future progress requires robust external validation, uncertainty-aware decision-making, interoperable infrastructure, cybersecurity, fail-safe control, and simultaneous optimisation of quality, safety, energy, productivity, and environmental and economic performance. The framework provides a foundation for benchmarking technological maturity and guiding the development of safe, scalable, and resource-efficient autonomous food-processing systems.