QR-DQN Robot Simulation Assets

This repository provides the simulation assets used in the paper:

"DESIGN A PATH – PLANNING STRATEGY FOR MOBILE ROBOT IN MULTI-STRUCTURED ENVIRONMENT BASED ON DISTRIBUTIONAL REINFORCEMENT LEARNING"

map1	map2	map3	map4

file_mapSmallAndDiffrb	file_mapSmallAndDiffrb2	file_mapSmallAndDiffrb3	file_mapSmallPersonAndDiffrb

file_DiffrobotAndMap	file_all	file_mapAndModelperson

Contents

src/
├── file_launch/                     # Launch files and map images
│   ├── CMakeLists.txt
│   ├── package.xml
│   ├── image_map/                   # Sample environment images
│   │   ├── file_all.png
│   │   ├── file_DiffrobotAndMap.png
│   │   ├── file_mapSmall_2AndDiffrb.png
│   │   ├── file_mapSmallAndDiffrb.png
│   │   ├── file_mapSmallPersonAndDiffrb.png
│   │   ├── small_1.png
│   │   ├── small_2.png
│   │   ├── small3.png
│   │   └── small4.png
│   └── launch/                      # ROS launch files
│       ├── file_all.launch
│       ├── file_DiffrobotAndMap.launch
│       ├── file_map.launch
│       ├── file_mapAndModelperson.launch
│       ├── file_mapSmallAndDiffrb.launch
│       ├── file_mapSmall_2AndDiffrb.launch
│       ├── file_mapSmall_3AndDiffrb.launch
│       ├── file_mapSmallPersonAndDiffrb.launch
│       ├── small1.launch
│       ├── small2.launch
│       ├── small3.launch
│       └── small4.launch
│
├── model_robot/                     # Robot model definition
│   ├── CMakeLists.txt
│   ├── package.xml
│   ├── meshes/
│   │   ├── base_link.STL
│   │   ├── cover_link.STL
│   │   ├── wheels/
│   │   │   ├── left_tire.stl
│   │   │   └── right_tire.stl
│   │   └── sensors/                 # Sensor meshes (LiDAR, camera, etc.)
│   │       ├── astra.dae
│   │       ├── HDL32E_base.dae
│   │       ├── HDL32E_base.stl
│   │       ├── HDL32E_scan.dae
│   │       ├── HDL32E_scan.stl
│   │       ├── hokuyo.dae
│   │       ├── kinectv2.dae
│   │       ├── kinectv2.stl
│   │       ├── kinectv2_nobase.dae
│   │       ├── kinectv2_nobase.stl
│   │       ├── lds.stl
│   │       ├── r200.dae
│   │       ├── VLP16_base_1.dae
│   │       ├── VLP16_base_1.stl
│   │       ├── VLP16_base_2.dae
│   │       ├── VLP16_base_2.stl
│   │       ├── VLP16_scan.dae
│   │       └── VLP16_scan.stl
│   └── urdf/
│       └── diffbot_realsize.urdf
│
└── mymap/                           # Simulation maps
    ├── CMakeLists.txt
    ├── package.xml
    ├── replace_path.py
    ├── control_model_person/         # Scripts for human models
    │   ├── control_personStanding.py
    │   ├── control_personWalking.py
    │   └── control_personWalking_mS3.py
    └── src/
        ├── map.world
        ├── map_small.world
        ├── map_small_person.world
        ├── map_modelperson.world
        ├── map_Small_2.world
        ├── map_Small_3.world
        ├── small_1.world
        ├── small_2.world
        ├── small3.world
        └── small4.world

---

Usage Instructions

1. Clone the Repository

Download the external models used in the simulation and place them in your home directory:

git clone https://github.com/phamduyaaaa/qr-dqn-robot-simulation-assets.git

2. Build the Workspace

From your catkin workspace root:

catkin_make

3. Adjust Device Username in Map Files

Some .world files contain absolute paths.
To update these paths to your local username:

Open mymap/replace_path.py.

Set the variable device_name to your system username.

Run the script:

python3 replace_path.py

4. Launch a Simulation Example

Use the corresponding launch file. For example:

roslaunch filelaunch file_map.launch

Notes

This repository provides simulation assets (SDF/World maps and URDF models) to ensure reproducibility.

The QR-DQN training and evaluation code is not included here but is available upon request from the corresponding author.

The external models folder is required for correct rendering of robots and environments.

Citation

If you use these assets in your research, please cite:

Nguyen, A.-T., et al. (2025). Design a Path–Planning Strategy for Mobile Robot in Multi-Structured Environment Based on Distributional Reinforcement Learning.

Contact

For questions or requests (including access to the training code and configuration files), please contact:

Nguyen Anh Tu – Corresponding Author

Email: tuna@haui.edu.vn

License

This repository is licensed under the MIT License.
You are free to use, modify, and distribute this work for research and educational purposes, provided that proper credit is given.