A novel friction model for hybrid robots considering configuration and velocity dependencies

Traditional friction models typically assume fixed joint loads and consider only the influence of velocity, which limits their generalizability under varying load conditions. In hybrid robots, however, joint loads vary with configuration. To address this issue, this study proposes a novel friction model for hybrid robotic systems that accounts for both joint velocity and configuration-dependent load variations. The model was experimentally validated on three parallel joints of the TriMule-200 hybrid robot across multiple configurations. The results demonstrate that the proposed model accurately captures the dual dependency of joint friction on velocity and configuration. When applied for friction compensation, the model reduced trajectory tracking errors by approximately 10% compared to traditional models, and by around 18% relative to the uncompensated case, confirming its effectiveness and robustness under diverse operating conditions.