Model Parameter Identification of a Machining Robot Using Joint Frequency Response Functions

The utilization of machining robots using industrial six degree-of-freedom (DOF) robot manipulators in the general cutting process is hindered by their cutting instability caused by a significantly lower dynamic stiffness than that of conventional machine tools. The dynamic stiffness of the machining robot is particularly low in low order eigenmodes of approximately 30 Hz or less compared to conventional machine tools. And it is also low in higher order eigenmodes, resulting in reduced cutting stability over the entire machining speed range. In this paper, we propose a frequency response function (FRF) estimation model using an 18 DOF multibody dynamics (MBD) model, which accounts for the key eigenmodes influencing low and high-speed milling, for the purpose of advancing stability improvement methods in robotic machining. First, a method for joint rotational FRF measurement in the joint frame for orthogonal three-directional excitations in the fixed frame is proposed. This not only enables quantitative evaluation of the influence of each joint on the structural dynamic behavior, but also enables model parameter identification with improved accuracy by considering the dynamic behavior of each joint. Subsequently, we employ particle swarm optimization (PSO) to identify the numerous parameters associated with the MBD model, which encompasses an implicit form of equation of motion. Finally, the proposed method is validated by comparing the estimated and measured FRFs.