Accurate characterization of the true tire–pavement contact state is essential for understanding pavement friction; yet conventional texture indicators and nominal contact assumptions cannot directly represent the actual interfacial interaction between rubber and pavement. This study proposes a rapid and non-destructive method for measuring three-dimensional tire–pavement true contact texture under different loads. A materials testing system was used to apply controlled loads to a rubber pad–carbon paper–pavement assembly, and the resulting imprints were combined with three-dimensional laser profilometer data and support-curve-based slicing to determine the real contact area ratio, penetration texture depth, and self-affine fractal dimension. Tests on nine asphalt pavement samples under loads from 5 to 20 kN showed that the real contact area ratio increased with load but remained below 40% at 20 kN. The predicted contact area from the reconstructed 3D texture agreed well with the imprint-based results, with an absolute error not exceeding 2.59%. Penetration texture depth showed a stronger relationship with skid resistance than fractal dimension. The proposed method provides a practical means of capturing effective tire–pavement contact parameters and offers useful inputs for laboratory-based skid resistance evaluation and texture-informed friction modeling.

Continuous eye tracking is difficult to deploy in production vehicles, while eye-movement-derived proxy labels are susceptible to same-source circularity and road-type confounding. We propose a driver-independent, confounder-aware cross-source framework in which the eye-movement-derived visual load index (VLI) proxy is constructed within training folds, whereas held-out prediction uses only vehicle kinematics and an optional driving-style prior. Evaluation used naturalistic data from 33 drivers, comprising 24,437 windows along a 48.4 km route, under driver-grouped cross-validation with driver-level bootstrap confidence intervals and control and increment tests. The vehicle-only model achieved an AUC of 0.589 across held-out drivers, rising to 0.633 with the style prior. After controlling for road type, vehicle features improved AUC by 0.042 (95% CI: 0.006–0.078), indicating measurable information beyond the binary road-type proxy; the style increment remained exploratory because its confidence interval included zero. Future-state prediction converged to a persistence baseline at 20 s. The VLI was not externally validated against an independent subjective or physiological criterion; accordingly, the findings are limited to cross-source prediction of an eye-movement-derived proxy. Across five architectures, this cross-source pattern was directionally consistent, supporting prospective evaluation of vehicle kinematics as a complementary sensing channel in multisource driver-monitoring systems.