Parameters Identification And Friction Compensation Based On The LuGre Model In Servo System

Friction exists in almost all of the mechanical motion. Friction is one of the main factorsaffecting the performance of servo system, which can lead to the tracking error, limit cycles,stick-slip and so on. Therefore, in order to compensate friction and meet the performancerequirements of the system, we must choose the friction model as close as possible to the realfriction model, and design the proper controller. In this paper, we research on the nonlinearfriction, which is based on the LuGre friction model. This dissertation focuses on the followingwork:First of all, the background of friction and research status are known by reading a lot ofrelated literature at home and abroad. In the mechanical rotation motion platform, the staticparameters of the LuGre friction model are identificated by curve fitting. Combined with thetest results, we compete the identification of the dynamic parameters.Secondly, using a dual observer to estimate the average deflection of the bristle in theLuGre friction model, we design a robust adaptive controller which is similar to the PIDadaptive control. The stability of the system is proved by selecting a suitable Lyapunovfunction. Compared with the simulation results of the PD control strategy, the system showsthe better tracking performance.Then, according with the neural network, the backstepping adaptive control is used tocompensate the nonlinear friction in servo system. We adopt the radial basis function toapproximate the external unmodeled disturbance. At the same time, the adaptive law isdesigned by the leakage. The control algorithm shows the transient performance andsemi-globally uniformly bounded stability of the system. And the inertia of the system and thereconstruction error are bounded. The simulation results verify the feasibility of the controlstrategy.Finally, the sliding mode surface is designed by the terminal switching function. Someactual assumptions are made, such as the unknown parameters and external disturbance arebounded. The states of the sliding surface can converge to the equilibrium point by theterminal switching function in finite time. The asymptotic stability of the closed-loop system isproved by stability theory. The simulation results verify the theoretical results. We mainly quantitatively analyzed on the transient performance of the system and obtain the boundness ofthe unknown parameters. Compared with the external disturbance and without disturbance, thefeasibility of the control strategy is proved. At the same time, the simulation results show theimpact of external disturbances.