GAO-Enhanced LADRC for a Tendon-Driven Compliant Joint Manipulator: Design, Modeling, and Control

This study presents a tendon-driven compliant joint (TDCJ) manipulator with a hybrid actuation design, where a 4-DOF tendon-driven elbow–wrist module is combined with a 2-DOF rigid shoulder joint module to reduce forearm weight and facilitate safer human-robot interaction. Although this hybrid design provides an efficient platform, the core contribution lies in integrating the Giant Armadillo Optimization (GAO) algorithm with Linear Active Disturbance Rejection Control (LADRC). This integration enables enhanced disturbance rejection and automatic parameter tuning in the TDCJ manipulator. The simulation results demonstrate that GAO-LADRC consistently outperforms conventional PID, Fuzzy PID, and LADRC controllers. Compared with a Particle Swarm Optimization (PSO)-based LADRC (PSO-LADRC), GAO-LADRC achieves comparable tracking performance at the elbow joint and superior accuracy and robustness at the wrist joint, while requiring less computational time during offline parameter optimization. These advancements make the TDCJ manipulator well suited for applications in home care robotics and other human-robot interaction scenarios, where compliance and precision are critical.