Digital twin-enabled cooperative autonomous truck platooning framework for modeling, cost estimation, and decision-making in transportation systems

Cooperative Autonomous Trucks (CATs) play a critical role in enabling the intelligence and automation of logistics. However, the involvement of multiple carriers, diverse autonomous trucks, and complex transportation environments presents substantial challenges to the large-scale deployment. This study proposes a Digital Twin (DT)-enabled framework for CAT to address these challenges, including transport task synchronization, platoon formation, and revenue estimation. The framework is structured with three interrelated components: (i) a Multiobjective Mixed Integer Programming (MOMIP) model for CATs platooning to optimize task allocation and decision-making, (ii) the DT-enabled kinematic and dynamic models for real-time synchronization and energy estimation, and (iii) real-time control and communication topologies to bridge optimization and practical implementation. A Resource Directive Decomposition (RDD) algorithm is introduced to efficiently solve the MOMIP model by decomposing the problem and applying disjunctive relaxation. Numerical validations show that the RDD algorithm achieves significantly faster convergence than the ϵ-constraint method when MOMIP has more than two objectives. CATs platoon formations achieve energy savings of approximately 10-14 % compared with historical human-driven trucks, and the proposed adaptive control with DT synchronization effectively ensures safety during emergency braking scenarios. An empirical validation using data from a real-world cross-border construction logistics project demonstrates cost savings and enhanced carrier compatibility.