Abstract

Understanding the stress-dependent behavior of unbound granular materials (UGMs) is essential for designing durable, resilient, and cost-effective transportation infrastructure. UGMs, including subgrades, subbases, and base layers, exhibit non-linear, anisotropic, and rate-dependent mechanical responses under traffic loading, which significantly influence pavement performance, structural integrity, and long-term serviceability. This paper presents a conceptual model for characterizing the stress-dependent behavior of UGMs, integrating constitutive modeling, laboratory characterization, and field calibration to predict strain, stiffness, and permanent deformation under variable loading conditions. The model emphasizes the interaction between vertical and lateral stresses, the evolution of resilient modulus with confining pressure, and the influence of stress path history on deformation and fatigue accumulation. The conceptual framework combines laboratory testing—such as repeated load triaxial, resonant column, and cyclic simple shear tests—with empirical and mechanistic relationships to develop predictive transfer functions for performance indicators including resilient modulus, permanent strain accumulation, and shear strength. Field calibration using non-destructive testing, instrumented pavements, and real-time traffic monitoring ensures that laboratory-derived parameters accurately reflect in-situ behavior, accounting for material heterogeneity, compaction variability, and moisture sensitivity. The model further incorporates probabilistic and reliability-based approaches to account for uncertainties in traffic loads, material properties, and environmental conditions, enabling robust predictions of long-term structural performance. By systematically characterizing the stress-dependent behavior of UGMs, the proposed model supports optimized pavement layer design, improved maintenance planning, and enhanced lifecycle performance. It also provides a foundation for integrating emerging technologies such as digital twins, machine learning, and automated materials testing to enable adaptive and predictive infrastructure management. Overall, the framework contributes to the development of resilient, sustainable, and cost-effective transportation systems capable of withstanding variable and uncertain loading conditions.