Abstract
Abstract
Friction, as a nonlinear and complex phenomenon, significantly affects the performance of mechanical systems requiring high-precision motion control and force feedback. Accurate modeling of frictional behavior is essential for effective control and compensation. The LuGre friction model is widely used due to its computational efficiency; however, it retains only the first-order displacement term, resulting in limited accuracy and noticeable drift, which restricts its use in precision applications. To address these issues, this study proposes an improved friction model based on the LuGre framework. Discrete bristles are introduced to incorporate the influence of surface topography at the contact interface. Additionally, an iterative numerical scheme is employed to enhance the computational accuracy. The simulation results demonstrate that the model captures key frictional phenomena, including stick–slip transitions, hysteresis, and friction lag. It also shows clear advantages over the LuGre model in representing non-local memory and non-drift characteristics. Experimental validation was conducted using a constant-velocity reciprocating test and a microamplitude sinusoidal excitation test, enabling stepwise parameter identification for slipping and sticking phases. A dual-frequency sinusoidal excitation test was further designed to evaluate the model performance under complex dynamic loading. The simulated friction forces agree well with the experimental measurements, verifying the model’s effectiveness and robustness. The proposed model enhances both accuracy and physical realism in friction modeling and can be applied in high-precision systems to solve the friction force output under the given input conditions.