YBDriving

Vision-based Autonomous Driving simulation among Unity virtual environment with Reinforcement Learning

Python Scripts: In the YBAutoDriving\Assets\Scripts\Python folder.

---------------------User Manual--------------------------

1. Hardware Requirements
   The “YB-Driving” can be executed with the following configuration:
   ------------------------------------------------!
   Part Recommended Configuration |
   ------------------------------------------------|
   CPU 12700H or above 10800H |
   Memory 32GB or above |
   Graphics Card Nvidia 3070Ti or above |
   Hard Disk NVME PCIE gen.4 NVME PCIE gen.3 |
   Network Not Required |
   OS Windows 11 |
   ------------------------------------------------|

2. Software Requirements
   Unity Engine Unity 2021 or newer (you may try Unity 2019 but not guranteed)
   The plug-in SDK “TMPro” is required.
   Python: Python 3.9 – 3.11
   Python IDE: VsCode
   Yolo Yolov8 is recommended, but lower version is ok

Environment Assets: Included in project solution files

4. Installation and execution
   Step1. Download the project files:
   

Step2. Use installed Unity Engine open the folder “YBAutoDriving” which contains Unity project.
Step3. Assign the output path of the training images:

Step4. Run the Unity project, and randomly click on the top-down map to set destination for the car. The qualified vision will automatically labelled and output to the set training image path.
Step5. Use your python IDE to open the “yolov8” folder, and run the “ybdriving.ipynb” script. Keep in mind to update the training image path. The yolo model will be trained.
Step6. Check the “runs” folder of the python project, and find the training result.
Step7. As an option, you may use yolo to test some image with traffic signs.

Copyright:
The project is opensource available at https://github.com/imadyTech/YBDriving under MIT license.
The copyright belong to the respective authors.

Credit:

Unity asset: SimplePoly City - Low Poly Assets:
https://assetstore.unity.com/packages/3d/environments/simplepoly-city-low-poly-assets-58899
Provided by VenCreations
www.vencreations.com/
[email protected]
https://assetstore.unity.com/publishers/19573
Under Unity Assetstore Standard EULA
https://unity.com/cn/legal/as-terms

2021 BMW M4 Competition:
Ricy
https://sketchfab.com/ngon_3d
https://sketchfab.com/3d-models/2021-bmw-m4-competition-d3f07b471d9f4a2c9a2acf79d88a3645

Traffic Signs:
Produced by Zhenqun, Shen with tripolygon_uModeler
[email protected]
Offered under MIT License