Build your own Autonomous Driving Car

Introduction

Enjoy the fun of your own autonomous driving car with the support of DPU， feel free to try your own AI model and new ideas.

Material

Download the image we provided here. To understand the whole project well, you can read the paper we published in FPT2019 here or the paper in MetroCAD2020 here. And here is one simple guide written in Chinese, Xilinx HydraMini试玩教程

Preparations

• The host pc should be ubuntu 16.04, you can refer to this to build an environment for DNNDK_v3.0 using tensorflow. If you want to run the simulator, refer to Autonomous-Driving-Simulator to set the Unity3d environment.
• Power the Pynq-Z2 in your car and run nmtui to connect the car to your switch(make sure your PC connected to the same).
• Read the guides in gitbook HydraMini Gitbook.

Learning Steps

1. Read Pynq Part/Collect Data to learn how to use your keyboard for car controlling. And then collect some images for training.
2. Read Host Part/Preprocess & Train and preprocess the collected images, then use the preprocessed images to train the model.
3. Read Host Part/Quantize & Compile and do quantization and compilation to the model.
4. Read Pynq Part/Run the Car and run the autonomous driving car.
5. If you want to try Autonomous Driving Simulator based on Unity3D, please refer to Virtual Part.