In order to reduce influence of friction on properties of robotic low-speed motion

generalized Maxwellslip（GMS） model is introduced which can describe characteristic of friction in low-speed stage. Parameters identification is realized by genetic algorithm and low-pass filter is implemented to filter high-frequency noise. To validate the effectiveness of this model

design trajectory tracking experiments on the manipulator

and GMS model is involved as feedforward compensation. Results of experiments show that GMS model can well describe the variety of friction in low-speed motion and further improve robotic tracking performance compared to widely used Stribeck friction model.