Material-Driven Stabilization of Tendon-Driven Compliant Actuators Through Cyclic Preconditioning
Abstract
Compliant actuators are widely employed in human–robot interaction and unstructured environments due to their inherent compliance and safety. However, the non-stationary mechanical behavior of compliant materials under cyclic loading often induces output drift and degraded motion repeatability, particularly in tendon-driven systems. This work proposes a material-driven stabilization strategy for compliant actuators through cyclic preconditioning of elastomeric materials. Inspired by biological muscle–tendon adaptation, the proposed approach exploits the material-state evolution of natural rubber to achieve a stable mechanical response with improved consistency. Experimental results show that the stress–strain response progressively converges toward a repeatable state after approximately eight loading cycles. A dual-network constitutive model is further developed to characterize the evolution of the permanent and physical subnetworks during cyclic deformation. Based on this principle, a tendon-driven compliant actuator integrating preconditioned natural rubber is developed, achieving improved open-loop motion repeatability without additional sensing or complex compensation strategies. Experiments demonstrate progressive trajectory convergence in a single-joint platform and a 39.83% reduction in trajectory RMSE in a biomimetic robotic arm. The proposed method provides a material-state-regulation perspective for improving the intrinsic stability of compliant robotic systems.