Genetic Algorithm for Parameter Identification of Free Section Friction

摘要

Friction system is one of the most familiar objects in mechanical system which has nonlinear characteristics. For precise-position and low-speed control, control strategy often relies on friction parameter identification. Considering the limitation of cantilever itself, its friction parameter identification is far more difficult than general. During the research in this paper, real experiment environment and virtual simulation are combined to enhance the understanding of whole mechanical system which is called Augmented Reality Method (ARM). But when ARM is applied to cognize the system motor process, it is found that traditional method which uses Coulomb Model (CM) is hard to satisfy precision requirement during the system-modeling stage. Because the object of study in this paper is a high-speed process, it is necessary to consider the impact of collision and vibration on cantilever. According to that, sine components are brought in to perfect CM which is called Augmented Coulomb Model (ACM) in order to get a more precise system friction model. After setting up ACM, Genetic Algorithm (GA) is used to realize the parameter identification of the cantilever system. The result points out that the identification precision of using ACM is more precise than CM for high-speed system modeling.