Cellulose-Based Gels for Intelligent Tactile Electronic Skin: From Materials Design to Dexterous Robotic Manipulation

Cellulose-based gels are emerging as promising soft materials for tactile electronic skin (e-skin) because of their tunable network structures, mechanical compliance, sustainability, and compatibility with ionic and electronic conduction. This review examines recent progress from material and structural engineering to tactile sensing, machine learning (ML)-assisted signal interpretation, and robotic manipulation. Particular attention is given to regenerated cellulose, nanocellulose, bacterial cellulose (BC), and carboxymethyl cellulose (CMC) systems, together with hierarchical porosity, anisotropic architectures, conductive fillers, and interfacial design strategies that improve sensitivity, sensing range, response speed, and mechanical stability. Representative demonstrations of wearable sensing, spatial pressure mapping, multimodal tactile acquisition, and sensor array integration are summarized and compared. The review further discusses how ML methods can extract task-relevant information from nonlinear and multimodal gel sensor signals, including handwriting, motion, material, texture, and object recognition. Studies on non-cellulosic hydrogel e-skin systems are included where they provide transferable approaches for tactile reconstruction, robotic integration, slip detection, and feedback control. Current evidence supports cellulose-based gels for soft tactile sensing, whereas transferable ML interpretation and robust robot-mounted, closed-loop manipulation remain less extensively validated. Future progress will require closer integration of material design, independent data validation, multimodal sensing, environmental stability, and task-specific robotic evaluation.