This replication package accompanies the manuscript “Compatibility-Invariance Conflict: a CLF-CBF surrogate safety measure for vehicle trajectories,” submitted to the Communications in Transportation Research under manuscript ID COMMTR-2026-0264. The code accompanying this manuscript is provided with the scripts, environment specification and instructions required to reproduce the reported analyses after the relevant source data are obtained under their original access conditions. SHRP2 reconstructed trajectories are publicly available from the source record at `https://doi.org/10.15787/VTT1/T7UUC1`. The reconstructed 100-Car files and source-data information are available from `https://github.com/Yiru-Jiao/Reconstruct100CarNDSData` and `https://doi.org/10.15787/VTT1/CEU6RB`. CIDAS source data cannot be publicly redistributed; authorized users must obtain the three documented cases through the applicable provider or institutional process.  The data/examples/ directory contains small public-derived fragments for loader validation only and does not​ satisfy the full input contract.

This code-only replication package corresponds to the revised manuscript “Vision Language Model-Guided Semantic Distillation Network for Pedestrian Crossing Intention Prediction” submitted to Communications in Transportation Research (Manuscript ID: COMMTR-2026-0077).

The package contains the implementation of the proposed VLM-GSDN framework, including scripts for generating pedestrian-level and scene-level semantic descriptions using Qwen, InternVL, MiniCPM, and Phi models, as well as training and evaluation workflows for the main model and different VLM-GSDN variants on the JAAD and PIE datasets. It also provides configuration files, environment specifications, and supporting documentation. A Replication Explanatory File describes the package structure, data preparation, model training, and evaluation procedures.

This package contains the numerical propagated interaction-field model, context-conditioned physics-informed neural network (PINN), field-aware reinforcement-learning pipeline, and processed measurements supporting the principal validation and simulator findings. 

The statistical workflow preserves the original observation/reward factorial and replaces only its complete-model merge candidate with the selected 0.25-times comfort profile. All five calibrated merging Soft Actor-Critic checkpoints and their five paired stock checkpoints are included. Policy-training seeds are 0-4; each policy is evaluated on traffic seeds 10000-10049 at density 0.3. The calibrated profile uses jerk weight 0.000375 and steering/throttle-change weights 0.00025. An indexed mapping connects each checkpoint to its effective training settings and episode CSV. 

The PINN checkpoint is accompanied by its available architecture, input channels, normalization, numerical-solver settings, loss weights, calibration-cache signatures, hardware record, training duration and 6,000-step loss history. 

This replication package accompanies the manuscript “CF-Drive: Candidate-conditioned interaction risk estimation for retrieval-augmented autonomous driving,” submitted to the Journal of Intelligent and Connected Vehicles under manuscript ID JICV-2026-0117. The package contains the source code, experimental configurations, log-disjoint fold definitions, processed candidate-level results, evaluation and plotting scripts, and a Replication Explanatory File for reproducing the principal risk-estimation, matched-shuffle, and utility-selection results. The original NAVSIM data are not redistributed. Users should obtain the source data from the official NAVSIM repository and comply with its applicable terms of use.

This replication package accompanies the manuscript "Hierarchical Neural-Scope Detector for Silent Fault Detection in Vision-Language-Action-Based Autonomous Driving," submitted to Communications in Transportation Research.

The package contains experimental code, processed data, detector models, feature specifications, evaluation configurations, sample-level predictions, and supporting results. The materials are organized into five modules covering OpenDriveVLA detector development, UniAD2 cross-architecture validation, independent validation, component ablation and threshold-sensitivity analysis, and runtime overhead evaluation.

These materials correspond to the experiments and results reported in the main manuscript, appendices, and supplementary information. The package includes README files and a replication guide describing the data, code, and procedures for reproducing the reported results.

The materials are also available at https://github.com/xlqy-114/Hierarchical-Neural-Scope-Detector.