Multilevel Motion Control of Cable-Driven Hyper-Redundant Manipulators

The cable-driven hyper-redundant manipulator (CDHRM) is characterized by a high number of joints, high degrees of freedom, and flexible motion, which pose challenges to its motion control. This article presents a novel multilevel motion control method for CDHRM based on the concept of joint angle-cable force cooperative optimization. First, for end-effector control, a new finite-time optimization method based on macro–micro decomposition is presented to solve the inverse kinematics problem. By utilizing a new key joints selection method, we reconstruct the high degrees of freedom of CDHRM, decomposing the macro CDHRM into several micro low-redundant manipulators with hybrid fixed and moving bases. The inverse kinematics of these manipulators is then solved by a finite-time dual neural network model with dynamic weight adjustment and convergence time constraints. The proposed method improves end-effector control accuracy and can handle multilayer constraints, which are challenging for traditional methods. Second, a new dual-channel proportional integral derivative (PID) control method that considers both cable tension and cable length is proposed, using dynamic interval constraints to regulate cable forces without stiffness calculations. This method simplifies control design, improves robustness, and reduces computational complexity compared to traditional stiffness-based approaches. It also addresses the coupling characteristics between the driving cables of different joints in the CDHRM. Finally, the proposed method is validated through simulations and experiments (see the attached video).