The article is devoted to the identification of friction in the joints of a 3-DOF manipulator with fluidic muscles. For the identification of friction, four different models were proposed, namely linear friction model, extended friction model, non-linear friction model and LuGre model. The values of the coefficients of the proposed models were estimated in the Matlab Simulink program environment based on the comparison of the measured and simulated curves of the angle of rotation of individual joints. The corresponding friction models with adjusted values of their coefficients were validated using two statistical indicators, namely Root Mean Square Error and Sum Square Error. The validation process consisted of comparing the measured and simulated course of the rotation angle of the joint using the corresponding estimated friction model. The aim of the validation was to compare the performance of individual friction models and to choose the most optimal variant for the dynamic model of the investigated system.
Science Journal 