Transitioning from Lab-Based to AI-Assisted Balanced Mix Design: Comprehensive Overview of Research,Development,and Future Perspectives

Abstract

Many states have made enormous research efforts to explore the feasibility of integrating balanced mix design (BMD) within their asphalt pavement programs. However, these research conclusions drawn from limited laboratory and field data may not be applicable to typical mixtures across other projects and states. Big data analytics, along with artificial intelligence (AI), is a widely accepted method to address the issue. This paper first did a literature review on the main research topics related to BMD and identified their deficiencies from a data adequacy perspective. Next, the research efforts in AI’s application on asphalt mixture performance and pavement condition prediction, and mix design optimization were reviewed. The successful uses of AI in asphalt mixture show great potential in overcoming shortcomings of lab-based BMD. Consequently, this study proposes an integrated AI-based big data analytics (AI-assisted) BMD framework and outlines future work to achieve this framework. The proposed future work includes establishing a comprehensive database, determining performance thresholds using big data analytics, developing an optimization-based BMD procedure with AI-based predictive pavement performance models, and determining quality control (QC)/quality assurance (QA) specifications using machine learning associated with probabilistic models. The framework not only determines mixture composition with balanced performance but also achieves time and economic savings and environmental effects during laboratory BMD and pavement construction.