From passive threshold triggering to proactive safety regulation: Model-free adaptive robust impedance control with fuzzy-quantized strategy for robotic retinal membrane peeling

Retinal epiretinal membrane (ERM) and internal limiting membrane (ILM) peeling, the definitive treatment for vision-impairing retinopathy, imposes extreme demands on submillimeter motion stability and micro-Newton-level force control. However, conventional fixed-force-threshold control strategies fail to meet these requirements: their passive, latency-prone intervention mechanism frequently causes unintended iatrogenic tissue damage during microsurgical manipulation. To reconcile the inherent conflict between strict force safety constraints for surgical safety assurance and surgeons’ intraoperative operational autonomy, this paper proposes a novel integrated control framework combining adaptive robust impedance control (ARIC) with a fuzzy-quantized safety strategy. Its three core innovations are: (1) an ARIC controller enabling accurate variable-force tracking without explicit prior environmental dynamic modeling; (2) a sigmoid-based continuous quantification factor for fine-grained surgical safety assessment and surgeon intent inference; (3) a dual-input fuzzy adaptation mechanism for chatter-free smooth impedance parameter tuning, shifting from passive threshold-triggered reaction to proactive cooperative safety regulation. Experiments on 3D-printed eye phantoms and ex vivo porcine eyes show that the proposed method reliably constrains peak interaction forces within the clinical safety range, significantly mitigates force overshoot, and enables smooth atraumatic membrane peeling. This work provides a validated safety-oriented solution for robot-assisted retinal microsurgery.