Dynamic modeling, identification, and optimization of a cable-assisted robotic system (CARS) for structural stiffness enhancement

This paper presents a dynamic modeling and identification framework for a Cable-Assisted Robotic System (CARS), developed to enhance the rigidity of serial industrial robots by adding tensioned cables. A robot model is coupled with elastic and dissipative cable contributions to derive a posture-dependent Cartesian frequency response function (FRF) of the combined robot-cable system. Cable stiffness and damping are experimentally identified through displacement-controlled cyclic tension tests, capturing the nonlinear and tension-dependent behavior of cables with different geometries. The robot joint stiffness and damping parameters are then identified based on experimental FRFs using particle swarm optimization (PSO) and are obtained locally for different robot postures within the workspace. The accuracy of the proposed model and identification approach is validated through impact hammer experiments performed under different pulley configurations and cable tensions, and it showed close agreement between simulated and measured FRFs, particularly at dominant modes. Finally, the validated model is used to optimize the locations of the cable-routing pulleys using a genetic algorithm (GA), showing that proper routing of the tensioned cables can further enhance the directional dynamic stiffness of the system.