Novel Variable Stiffness Bionic Parallel Wrist: Synthesis, Modeling, and Analysis

Abstract This article presents a novel bionic wrist driven by twisted and coiled polymer (TCP) actuators, mimicking the anatomy of the human wrist, especially the carpal bones and motion-related muscles. First, equivalent fitting models are proposed to represent the radial–ulnar and flexion–extension joint axis motion trajectories. A type synthesis method for parallel wrists (PWs) matching these models is introduced using the Atlas method, leading to the synthesis of four novel PWs. Additionally, variable-stiffness TCP actuators are integrated on both sides of the PW, resulting in a bionic wrist that combines artificial muscles and joints. A general kinematic model of the TCP-driven PW is established based on the kinematic form of linear actuator equivalent joints, followed by workspace, singularity, and performance analyses. Moreover, a stiffness model for the PW is developed by fitting the TCP displacement–stiffness data. Using this model, the variable stiffness characteristics under unilateral and compound PW driving conditions are analyzed. Finally, prototype testing is conducted to validate the PW's trajectory and stiffness. This study serves as a reference for applying TCP-like linear smart actuators in parallel mechanisms.