Control Techniques for Friction Compensation

Abstract

Friction is a natural phenomenon that affects almost all motion. It has been the subject of extensive studies for centuries, with the main objectives being the design of effective lubricating processes and the understanding of the mechanisms of wear. Whereas friction effects at moderate velocities are somewhat predictable, it is the effect of friction at low velocities that is very difficult to model. The facts that friction changes sign with velocity, is asymmetric about the velocity axis, has evolutionary characteristics, and exhibits the stick-slip phenomenon, etc., aggravates the problem. Although friction effects have been well understood qualitatively, researchers have often relied on experimental data to formulate various mathematical models. A heuristic model for friction was first proposed by Leonardo da Vinci [11] in 1519; however, the model failed to capture the low-velocity friction effects such as the Stribeck effect, presliding displacement, rising static friction, etc., which play a major role in high-precision position/velocity tracking applications. In recent years, several dynamic models have been introduced to describe this highly nonlinear behavior exhibited by friction. For example, Dahl [12] proposed a dynamic model to capture the spring-like behavior during stiction. Canudas et al. [9] proposed a dynamic state-variable model to capture friction effects such as the Stribeck effect, hysteresis, spring-like behavior of stiction, and varying breakaway force.