Research And Experiment Of Friction Identification And Compensation In Mechatronic Systems

The friction is one of the main factors that influence mechanical servo controlsystem’s performance. How to accurately identify friction model and to implementcontrol compensation becomes a primary technology in reducing the friction effect.Based on typical mechatronic servo system, this paper first applies adaptiveidentification method and derives the adaptive identification implementation process ofthe coulomb+viscous friction model, coulomb+viscous+Stribeck friction modeland LuGre friction model. Then the identification model is used to completefeed-forward compensation. The corresponding simulation and experiment conductedshow the effectiveness of the method. After these, aiming at reducing the limitations ofmathematical modeling application, the fuzzy modeling method is proposed. Thismethod makes use of the fuzzy modeling method to construct the friction model. Also,the subsequent simulation and experiment compares quality of both the fuzzymodeling method and the adaptive friction identification method.Because the friction model in actual system is unknown, how to choose a suitablefriction model to implement compensation is the key to improve performance. Basedon ellipsoid unfalsified control theory, this paper puts forward the method of how tochoose the appropriate friction model according to requirements of systemperformance, in which classic mathematical model is used. Based on the typical springsports system, simulation is conducted to choose the friction model as well as to verifythe validity of the method.Based on IMDU rotation servo experimental platform and in view of friction,different identification and compensation methods are tested by system experiments, inpurpose of verifying the correctness of the above method.