Robustness study of the bio-inspired musculoskeletal arm robot based on the data-driven iterative learning algorithm
Jianbo Yuan, Jing Dai, Yerui Fan, Yaxiong Wu, Yunpeng Liang, Weixin Yan
TL;DR
This paper tackles robust, human-like manipulation with a musculature-inspired robotic arm by integrating a seven-DOF LTDM-Arm (with 15 modular artificial muscles) and a data-driven iterative learning control (DDILC) strategy. The authors employ a Hill-type muscle model to capture activation dynamics and muscle–tendon forces, and derive a comprehensive nonlinear dynamic model linking muscle excitations to joint motion through Jacobian-based mappings. DDILC leverages a biased-format dynamic linearization along the time axis to produce a data-driven linear surrogate (plus online updates for PJM, gain matrices, and feedforward terms) and achieves convergence of tracking error to near zero under repetitive tasks, including disturbances. Numerical simulations and prototype experiments show that the DDILC control yields substantially smaller trajectory errors than baseline methods (e.g., CMC, MTIS-FCC, Nonlinear-PID), with robustness to load disturbances up to $20\%$ in simulation and $15\%$ in experiments, and end-effector errors dropping to around a few millimeters after iterations. The work demonstrates the practicality of musculoskeletal-inspired robotics for robust, human-like dexterity in unstructured environments and provides a foundation for safe human–robot interaction and multi-sensor fusion in future tasks.
Abstract
The human arm exhibits remarkable capabilities, including both explosive power and precision, which demonstrate dexterity, compliance, and robustness in unstructured environments. Developing robotic systems that emulate human-like operational characteristics through musculoskeletal structures has long been a research focus. In this study, we designed a novel lightweight tendon-driven musculoskeletal arm (LTDM-Arm), featuring a seven degree-of-freedom (DOF) skeletal joint system and a modularized artificial muscular system (MAMS) with 15 actuators. Additionally, we employed a Hilly-type muscle model and data-driven iterative learning control (DDILC) to learn and refine activation signals for repetitive tasks within a finite time frame. We validated the anti-interference capabilities of the musculoskeletal system through both simulations and experiments. The results show that the LTDM-Arm system can effectively achieve desired trajectory tracking tasks, even under load disturbances of 20 % in simulation and 15 % in experiments. This research lays the foundation for developing advanced robotic systems with human-like operational performance.