Dynamic Resequencing for Autonomous Vehicle Coordination

14 concepts17 papers1 proposal

In traditional traffic management, the rule of thumb is simple: first come, first served. Whether it is a four-way stop or a highway merge, we generally expect the order of arrival to dictate the order of passage. This First-In-First-Out (FIFO) logic is computationally easy to implement and intuitively "fair," but as research into Connected and Automated Vehicles (CAVs) matures, it is increasingly viewed as a major bottleneck to systemic efficiency.

The FIFO Bottleneck

The fundamental flaw of FIFO scheduling is that it ignores the physical state of the vehicles involved. As Arnaud de La Fortelle notes in Analysis of optimal solutions to robot coordination problems to improve autonomous intersection management policies, strictly chronological policies fail to account for global efficiency. A single vehicle arriving slightly earlier on one path might force a large, high-speed platoon on a conflicting path to decelerate sharply, wasting energy and reducing throughput.

This lack of coordination creates what Au et al. describe as a failure to optimize "crossing windows." In their work on Enforcing Liveness in Autonomous Traffic Management, the authors demonstrate that rigid FIFO can lead to "starvation," where vehicles on side roads are indefinitely blocked by main-road traffic. By moving to batch processing—grouping requests rather than handling them one-by-one—they achieved up to a 3.6x reduction in average delay.

Expanding the Solution Space

When we relax the requirement that vehicles must exit an intersection in the exact order they arrived, the mathematical "solution space" for traffic flow expands significantly. Yue Zhang argues in A Decentralized Optimal Control Framework for Connected Automated Vehicles at Urban Intersections with Dynamic Resequencing that strict sequencing prevents systems from exploiting available gaps. By allowing "line-jumping"—where a closer or faster vehicle is prioritized over a distant arrival—Zhang demonstrated a 34% reduction in travel time.

This shift transforms traffic management from a simple queuing problem into a sophisticated scheduling problem. Andreas A. Malikopoulos draws on job-shop scheduling theory in A Priority-Aware Replanning and Resequencing Framework for Coordination of Connected and Automated Vehicles to treat each vehicle as a "job" that can be reordered to maximize the "factory" throughput.

Flexible Merging and Platooning

The benefits of dynamic resequencing extend beyond intersections to highway bottlenecks. Traditional merging often relies on pre-assigned sequences, forcing vehicles to make drastic speed adjustments to reach a specific slot. Hu et al. address this in C2TE: Coordinated constrained task execution design for ordering-flexible multi-vehicle platoon merging, proposing an algorithm that allows vehicles to merge from multiple lanes without a pre-set order. This flexibility ensures that if a vehicle breaks down or a new one joins, the system doesn't collapse under a rigid, outdated plan.

Similarly, in mixed-traffic scenarios where automated vehicles share the road with humans, fixed sequences are often impractical. Chang Liu points out in Decentralized On-Ramp Merging Control of Connected and Automated Vehicles in the Mixed Traffic Using Control Barrier Functions that fixed merging sequences prevent faster CAVs from safely overtaking slower human-driven vehicles, limiting the overall flow.

Bi-Level Optimization

To make dynamic resequencing computationally feasible, researchers often employ a bi-level architecture:

  1. Upper Level (Scheduling): A high-level coordinator decides the optimal sequence (who goes when).
  2. Lower Level (Control): Individual vehicles calculate the most energy-efficient trajectory to meet that schedule.

This approach, championed by Ye et al. in their Bi-level Optimal Edge Computing Model, allows for real-time adjustments. By decoupling the "ordering" from the "driving," the system can achieve massive gains—up to 60% energy reduction in some merging scenarios—while maintaining safety through Control Barrier Functions (CBFs), as explored by Xiao et al.

Beyond the Road

The limitations of FIFO are not unique to cars. In Data-Driven Optimization for Air Traffic Flow Management with Trajectory Preferences, Lancia et al. show that traditional air traffic heuristics fail to account for the "ripple effects" of scheduling. By moving away from simple arrival-order sequencing and using historical data to optimize flight paths, they reduced network delays by nearly 70%.

Whether in the air or on the ground, the research direction is clear: the most efficient path forward is rarely a straight line formed by the order of arrival. By embracing dynamic resequencing, autonomous systems can move past the chronological constraints of the past to unlock the full capacity of our infrastructure.

Key Foundational Work

Analysis of optimal solutions to robot coordination problems to improve autonomous intersection management policies (2016) — contributed 1 concept elements to this direction.
Enforcing Liveness in Autonomous Traffic Management (2011) — contributed 1 concept elements to this direction.
History & citationVersion 1 · 2026-10-04

Current · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_fifo_bottleneck_0a662f1c,
  title = {FIFO bottleneck},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/fifo-bottleneck/rdr_0a662f1c4f514d4e97690aea1081111e},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-19},
}

This exact version (version 1, frozen text):

@misc{lacuna_fifo_bottleneck_0a662f1c_v1,
  title = {FIFO bottleneck},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/fifo-bottleneck/rdr_0a662f1c4f514d4e97690aea1081111e/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-19},
}
About

These directions overlap with this one through shared papers.

See something wrong? Report a mistake