Efficient Identification and High-Fidelity Extrapolation of the Parameters in LuGre Frictional Model at Rubber–Metal Interface Utilizing Machine Learning and Physics-Informed Prior Model With Limited Experimental Dataset

Accurate and efficient prediction of friction behavior at rubber–metal interfaces is essential for advanced robotic and mechanical systems. However, it remains challenging due to nonlinear friction characteristics and costly experimental data. This article proposes an efficient and cost-saving framework that integrates machine learning, an optimization algorithm, and a physics-informed prior model for parameter identification and extrapolation of the LuGre model within limited experimental data. A hybrid Gaussian process regression (GPR)-particle swarm optimization method was developed to identify dynamic parameters within a limited normal load range (2–10 N). A physics-informed prior model was developed for parameter extrapolation, integrating Greenwood–Williamson contact theory with GPR to ensure physically consistent parameter evolution under unmeasured conditions. A tribo-pair featuring styrene-butadiene rubber/natural rubber block sliding against the aluminum plate was employed for verification on a block-on-disk test rig. The results demonstrated that the identified friction responses agreed well with measurements, achieving R2 values above 0.91 and normalized root mean squared error (NRMSE) values below 12%. Even with a 50% extrapolation beyond the training load range, the proposed framework maintained high predictive accuracy (R2 > 0.93, NRMSE < 10%). This study provides an efficient and physically interpretable methodology for friction parameter identification and extrapolation under limited-data conditions.