Intermodal freight terminals are critical nodes in North American supply chains, yet they face persistent congestion, long truck dwell times, and significant local air pollution. Major rail operators manage high-volume terminals where stochastic truck arrivals and limited coordination with yard operations lead to inefficiencies.
The rapid advancement of autonomous modular vehicle technologies (AMVT) creates an unprecedented opportunity: unlike human-driven vehicles, AMVT-based drayage fleets can adjust load capacity, departure times, speeds, and routes dynamically in response to terminal conditions. This controllability enables real-time coordination with rail yard crane operations and train schedules.
In Phase I of this project (5/1/2026 – 5/1/2027), the research team will focus on data collection, including drayage vehicle and train operational data as well as interviews with rail yard operating officers to understand the current state of the practice, challenges and opportunities (e.g., adopting advanced vehicle technologies such as AMVT) in railyard planning and operational decisions. Realistically we target to interview 2-5 railyards across the country. To supplement the small sample size, we plan to use LLM to help generate additional data for railyard operations and technology adoption profiles based on size and geographic location.
In Phase II, the research team will develop an integrated AI-based optimization framework to synchronize AMVT-based drayage operations with rail terminal processes, with the goal of reducing congestion and operating costs. Specifically, A bi-level optimization framework will be developed:
· Upper Level: Autonomous truck dispatch and routing optimization to minimize queueing delay, fuel consumption, and emissions.
· Lower Level: Rail yard operations optimization including crane scheduling, container stacking, and train loading/unloading sequences.
The model will incorporate mixed-integer linear programming (MILP) and reinforcement learning for adaptive control under uncertainty (e.g., demand and financial uncertainty).
This research improves railyard efficiency through integrated dispatch with rail yard scheduling, increased terminal throughput, and reduced truck dwell time by adopting emerging autonomous modular vehicle technology (AMVT).