This replication package accompanies the manuscript entitled "Vehicle Platooning and Signal Coordination for Intelligent Connected Intersections Toward Autonomous Transportation Systems," submitted to the Journal of Intelligent and Connected Vehicles (JICV). It provides SUMO-based simulation code, experimental configurations, processed traffic-demand data, prediction assets, and analysis scripts for reproducing the reported results. The code implements a rolling coordination framework integrating vehicle platooning, short-term traffic demand prediction, and signal timing optimization at an isolated intelligent connected intersection. The package supports comparative experiments under three traffic demand levels (1.0, 1.2, and 1.5) and includes the corresponding experimental results. The README provides software requirements, environment setup instructions, experimental configurations, execution commands, and procedures for regenerating the performance metrics, tables, and figures reported in the manuscript, including system delay, queue accumulation, throughput, and traffic clearance.

Replication materials

This repository provides the VTF-Flow replication package, including source code, experimental configurations, preprocessing, training and evaluation scripts, and reference results. The study uses the publicly available RELLIS-3D dataset, which can be obtained separately from the original provider.

Replication package associated with the manuscript “Towards Lawful Autonomous Driving: Deriving Scenario-Aware Driving Requirements from Traffic Laws” (Communications in Transportation Research, Manuscript ID: COMMTR-2026-0168.R3). The package includes the law corpus, source code, and README document to reproduce the numerical experiments.

ScenePilot-4K is a large-scale first-person driving dataset and benchmark for vision-language models. It contains 3,847 hours of video and 27.7 million front-view frames spanning 63 countries and regions and 1,210 cities. ScenePilot-Bench evaluates scene understanding, spatial perception, motion planning, and GPT-based semantic alignment.

This replication package contains the public code, README.md, and an explanatory file. Annotations, VQA datasets and evaluation subsets, video-level train/validation/test splits, source-video indices, and model checkpoints are available on Hugging Face. Download locations and usage instructions are provided in README.md. Original videos and extracted frames are not redistributed.