A physics-informed coupled back-calculation method for asphalt pavement modulus evaluation using TSDD measurements

Pavement modulus back-calculation suffers from non-uniqueness, instability and heavy computational loads caused by interlayer parameter coupling and insufficient response sensitivity. Targeting Traffic Speed Deflection Device (TSDD) measurements, this study develops a physics-informed coupled back-calculation framework for asphalt pavements using equivalent elastic deflection basins converted from TSDD signals. Finite-element simulations were first implemented to analyse responses of flexible and semi-rigid pavements under moving TSDD loads. TSDD measurements were converted into equivalent elastic deflection basins via an existing elastic conversion method to facilitate multilayer elastic analysis. Correlations between subgrade modulus and inertial-point parameters were established to build a deflection basin regularity (DBR) procedure for uniquely identifying subgrade modulus. This fixed subgrade modulus was adopted as a physical constraint in a Simulated Annealing–Particle Swarm Optimization (SA-PSO) matching algorithm to iteratively retrieve surface and base moduli, constructing the integrated DBR-Matching coupled framework. Numerical validation confirms favourable accuracy and robustness. All layer moduli of flexible pavements exhibit average relative errors below 10%. For semi-rigid pavements, the method surpasses conventional DBR and SA-PSO, with average relative errors of 2.59%, 6.14% and 12.18% for surface, base and subgrade moduli. Field TSDD verification shows stable convergence, deflection basin fitting RMSE mostly under 10%, and back-calculated moduli within reasonable engineering ranges. This framework integrates physical constraints and intelligent optimization, offering an accurate, efficient tool for TSDD-based pavement evaluation and network-scale assessment.