A review of falling weight deflectometer backcalculation methods for pavement moduli prediction

Abstract

This paper presents a comprehensive review of the major approaches used to estimate pavement layer moduli from Falling Weight Deflectometer (FWD) data, covering static, dynamic, and machine learning based backcalculation methods. Commonly used static backcalculation tools are examined based on their computational procedures, convergence characteristics, and suitability for different pavement structures. Recent developments in dynamic backcalculation are also discussed, highlighting their analytical frameworks, potential advantages, and practical challenges. The review further summarizes the growing role of machine learning, including neural networks, genetic algorithms, hybrid models, and probabilistic techniques, in improving prediction accuracy and reducing dependence on initial seed values. The strengths and limitations of these approaches are compared, and the capabilities and constraints of current tools are evaluated. In summary, this study brings together key findings from previous research, identifies important research gaps, and guides the development of user-friendly backcalculation methods for pavement evaluation and long-term infrastructure management.