Category
Data type
Tag
Region
MetaSSC: Enhancing 3D Semantic Scene Completion for Autonomous Driving through Meta-Learning and Long-sequence Modeling
Our approach begins with a voxel-based semantic segmentation (SS) pretraining task, designed to explore the semantics and geometry of incomplete regions while acquiring transferable meta-knowledge. Using simulated cooperative perception datasets, we supervise the training of a single vehicle’s perception using the aggregated sensor data from multiple nearby connected autonomous vehicles (CAVs), generating richer and more comprehensive labels. This meta-knowledge is then adapted to the target domain through a dual-phase training strategy—without adding extra model parameters—ensuring efficient deployment. To further enhance the model’s ability to capture long-sequence relationships in 3D voxel grids, we integrate Mamba blocks with deformable convolution and large-kernel attention into the backbone network. Extensive experiments show that MetaSSC achieves state-of-the-art performance, surpassing competing models by a significant margin, while also reducing deployment costs.

Road Network Data in Wangjing District, Beijing, China
‘NodeSet_2D.mat’ and ‘LinkSet_2D.mat’ describe the node set and link set in Wangjing road network, and ‘ODset.mat’ describe the O-D pairs of demand for UAVs.

SAFER-Predictor
This study utilizes datasets from the original baselines, including Los-loop & SZ-taxi (https://github.com/lehaifeng/T-GCN), PEMS-BAY and METR-LA (https://github.com/liyaguang/DCRNN), and PeMSD4 and PeMSD8 (https://github.com/wanhuaiyu/ASTGCN#datasets). The code is available at https://github.com/Chengfeng-Jia/SAFER-Predictor.

Replication package of the Paper titled a multi-agent social interaction model for autonomous vehicle testing
All the materials source coding and data of the Paper titled a multi-agent social interaction model for autonomous vehicle testing

Towards robust motion control in multi-source uncertain scenarios by robust policy iteration
Replication Package for "Towards robust motion control in multi-source uncertain scenarios by robust policy iteration"

Safety Assurance Adaptive Control for Modular Autonomous Vehicles
This study proposes a Safety Assurance Adaptive Model Predictive Control (SAAMPC) framework to achieve distributed docking/undocking operations for MAVs in uncertain environments. The SAAMPC framework integrates a Model Predictive Control (MPC) controller for trajectory optimization, an adaptive module for dynamic adjustment of control parameters with disturbance, and an adaptive safety assurance module with longitudinal and lateral Control Barrier Functions to ensure safe operation during risky and uncertain conditions. The effectiveness of the proposed approach is validated through simulations in Simulink and field tests on a reduced-scale MAV platform. Experimental results validate that the SAAMPC framework successfully ensures smooth and safe vehicle following and robust execution of docking/undocking operations under uncertainties.

ProChunkFormer
Replication Package for "Enhanced trajectory reconstruction from sparse and noisy GPS data: A progressive chunked transformer approach"

Replication package of the Paper titled a trajectory planning and tracking method based on deep hierarchical reinforcement learning
All the materials source coding and results coding of the Paper titled a trajectory planning and tracking method based on deep hierarchical reinforcement learning

Replication package of Enhancing Driver Emotion Recognition Through Deep Ensemble Classification.
All the materials source coding and results coding of Enhancing Driver Emotion Recognition Through Deep Ensemble Classification.

Extracting networkwide road segment location, direction, and turning movement rules from global positioning system vehicle trajectory data for macrosimulation
The replication package dataset (OpenstreetMap study network, GPS raw data, and intersection locations) used for the study.

Project: A Risk-based Unmanned Aerial Vehicle Path Planning Scheme for Complex Air-Ground Environments
Overview
It is a Python-based application aimed at designing and optimizing air corridors for efficient air traffic management. This project leverages various third-party risks to create safe and efficient air routes.
Features
- Route Optimization
- Safety Analysis
- Data Visualization
Installation
To get started with the Air Corridor Design project, follow these steps:
1. Download the data set.
2. Navigate to the project directory:
cd ./
3. Install the required dependencies:
pip install -r requirements.txt
Usage
To run the application, use the following command:
python ./Air_Corridor_Design/main.py
Project Structure
- README.md: This file.
- requirements.txt: List of dependencies required for the project.
- setup.py: Script for completing setup.
- Air_Corridor_Design/: Directory containing the source code.
- Data/: Directory containing data files used by the application.
- Experiments/: Directory containing experimental results.
- Scripts/: Directory containing some scripts for testing.
Contact
For any questions or suggestions, please contact [zhouk23@mails.tsinghua.edu.cn].





Traffic oscillations mitigation with physics-enhanced residual learning (PERL)-based predictive control
This repository supports the research on mitigating traffic oscillations using a Physics-Enhanced Residual Learning (PERL)-based predictive control approach. It contains all necessary components related to both the prediction and control aspects of the study. The prediction module includes the pre-processed NGSIM dataset, prediction models, and the resulting predictions, which focus on forecasting the behavior of preceding vehicles, including speed fluctuations, to allow timely responses. The control module implements a Model Predictive Control (MPC) approach that uses the prediction results to control connected and automated vehicles (CAVs), enhancing safety and comfort in mixed traffic environments. All code, data, and results are included to ensure that users can replicate the experiments and validate the findings effectively.

Bidirectional Q-Learning for recycling path planning of used appliances under strong and weak constraints
The Layered Bidirectional Q-Learning (LBQ) algorithm is designed for path planning, tackling the complexities inherent in multilayer path planning during the recycling process. This approach incorporates a bidirectional update mechanism that minimizes the unpredictability associated with initial exploration phases. Additionally, the algorithm employs a hierarchical reinforcement learning strategy, which breaks down intricate tasks into more manageable subtasks. Through the strategic design of reward functions that address various constraints, the LBQ algorithm successfully optimizes paths under multiple conditions.

Collaborative electric vehicle routing with meet points
The replication file contains data used in the paper "Collaborative electric vehicle routing with meet points" published in COMMTR. In this paper, we use real-world locations of grocery stores in Gothenburg, Sweden. The original data is the real addresses. Additionally, we use some test location data to evaluate large-scale instance performance, with customer locations randomly generated within a 25 km × 25 km region. The x and y coordinates range from 0 to 25 and the unit is kilometer. In addition to customer locations, we have included meet point and depot locations in the file. The type of each location and its associated company are noted. This data has been prepared for replication purposes.
The code for the proposed algorithms is currently being used in another paper, which is under review. We will upload the code once the other paper is published.

Replication package for the article "Bridging the gap: towards a holistic understanding of shared micromobility fleet development dynamics"
This repository provides the system dynamics simulation model to the paper titled "Bridging the gap: towards a holistic understanding of shared micromobility fleet development dynamics", submitted to the Journal Communications in Transportation Research. The simulation model is made available as a Vensim Packaged Model in the file format vpmx. The free software Vensim Model Reader is necessary to view and simulate the model (https://vensim.com/free-downloads/#Model_Reader). The Model Reader allows read-only access to models created with Vensim. Policies and policy combinations can be defined and simulated using the slider controls. All equations and parameter values can be inspected using the “Document” and “Document all” tools. It furthermore includes the complete simulation results for all parameters and variables for ten different policy packages in the file format csv (Comma Separates Values).

Segmented trust assessment in autonomous vehicles via eye-tracking - dataset to JICV manuscript
During the eye-tracking measurement—using a Tobii Pro eye camera—the participants were asked to watch a video shot inside an autonomous vehicle. The video was compiled from publicly available sequences according to two aspects:
1) Different situations are shown from different positions,
2) The eye movements observed at the beginning and at the end of the video can be compared to some extent; therefore, we selected the first and last sections of the video in such a way that the subjects can watch the ride from the same angle and have relatively more time to become involved in the given situation.
The recording consisted of 5 different sequences, within which we defined 29 areas of interest (AOI)
In terms of the data collected via eye cameras, four important measures can be defined:
− Average fixation duration: how long a fixation lasted on average within the given AOI.
− Fixation count: number of fixations within the given AOI.
− Total fixation duration: total length of fixations within the given AOI.
− Pupil diameter (right/left): the change in right and left pupil diameter within the given AOI.

RAAML of urban intersection cyber physical system architecture design
- This dataset contains all the RAAML files for the architecture design of the urban intersection information physical system proposed in this studyTo open this file, the MagicDraw software mentioned in this article needs to be used.
- The architecture design file utilizes the RAAML database released by OMG by extending the functionality of UML and SysML. This database includes basic models related to reliability analysis, which have now been imported into the provided dataset. Please refer to OMG Risk Analysis and Assessment Modeling Language, Risk Analysis and Assessment Modeling Language (RAAML) Libraries and profiles for the original file, Version 1.0, OMG Document Number: ptc/21-12-02, http://www.omg.org/spec/RAAML.
- The data involved in Section 4.6 of this article are all the results obtained in Section 4.5, and the response surface method built-in in Design Expert is used for relevant calculations.

VTOL sites location considering obstacle clearance during approach and departure
This dataset constains data and codes for determining the VTOL sites.

Empowering highway network: optimal deployment and strategy for dynamic wireless charging lanes
The OD data in this paper are from the real road network data of Guangdong province. The initial SoC in the paper is randomly generated and follows a normal distribution.

Multi-Level Objectives Control of AVs at A Saturated Signalized Intersection with Multi-Agent Deep Reinforcement Learning Approach
Code and experimental data for "Multi-Level Objectives Control of AVs at A Saturated Signalized Intersection with Multi-Agent Deep Reinforcement Learning Approach."Codes are written in Python and developed on an open-source tool “Flow”.( https://flow.readthedocs.io/en/latest/index.html)
