OASim

KO | EN

gitlite — search

#python #python3 #nodejs #linux #docker #react #react-native #blockchain #cryptocurrency #laravel #typescript #mongodb

★ 247 Open GitHub ↗

OASim: an Open and Adaptive Simulator based on Neural Rendering for Autonomous Driving

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# OASim: an Open and Adaptive Simulator based on Neural Rendering for Autonomous Driving

[![Custom badge](https://img.shields.io/badge/Paper-Arxiv-b31b1b?logo=arxiv&logoColor=white?style=flat-square)](https://arxiv.org/abs/2402.03830)
[![Custom badge](https://img.shields.io/badge/Project%20Page-white?logo=GitHub&color=green?style=flat-square)](https://pjlab-adg.github.io/OASim)
[![Stars](https://img.shields.io/github/stars/pjlab-adg/OASim?style=social)](https://github.com/pjlab-adg/OASim/stargazers)
[![GitHub issues](https://img.shields.io/github/issues/PJLab-ADG/OASim?style=flat-square)](https://github.com/PJLab-ADG/OASim/issues)
[![PRs Welcome](https://img.shields.io/badge/PRs-welcome-brightgreen.svg?style=flat-square)](https://github.com/PJLab-ADG/OASim/pulls)

https://github.com/PJLab-ADG/OASim/assets/24562717/ddb580c4-d329-419e-87f2-f3c8f487d8f7

## 🔍 Framework Overview

OASim focuses on generating new and highly customizable autonomous driving data through neural implicit reconstruction and rendering techniques. This technology has a wealth of applications, such as large-scale data and scene generation, corner case generation, autonomous driving closed-loop training, autonomous driving stack testing, etc.

## 🌟 Highlights
- **`2024-02-08`** Codes are now release!
- **`2024-02-07`** Explore our project page, now live [here](https://pjlab-adg.github.io/OASim)🔗!
- **`2024-02-07`** Our paper is available on [Arxiv](https://arxiv.org/abs/2402.03830)📄!

## 🚀 Getting Started

First, clone with submodules:

```shell
git clone git@github.com:PJLab-ADG/OASim.git --recurse-submodules -j8
```
### 1. Installation 📦

Our code is developed on Ubuntu 22.04 using Python 3.9 and PyTorch 2.0 (CUDA 11.8). Please note that the code has only been tested with these specified versions. We recommend using conda for the installation of dependencies. The installation process might take more than 30 minutes.

```bash
conda create -n oasim python=3.9 
conda activate oasim
conda install pytorch==2.0.0 torchvision==0.15.0 torchaudio==2.0.0 pytorch-cuda=11.8 -c pytorch -c nvidia
```
- pytorch_scatter

```shell
conda install pytorch-scatter -c pyg
```

- other pip packages

```shell
pip install opencv-python-headless kornia imagesize omegaconf addict imageio imageio-ffmpeg scikit-image scikit-learn pyyaml pynvml psutil seaborn==0.12.0 trimesh plyfile ninja icecream tqdm plyfile tensorboard PySide6 vtk dash-vtk numpy==1.24.2 dearpygui==1.8.0 matplotlib==3.7.1 pandas==2.0.0 pynput==1.7.6 rich==13.4.2 sumolib==1.16.0 traci==1.16.0
```
- the most recent sumo

```shell
sudo add-apt-repository ppa:sumo/stable
sudo apt-get update
sudo apt-get install sumo sumo-tools sumo-doc
```

### 2. Configuration ⚙️

`cd` to the `neuralsim/nr3d_lib` directory.

Then: (Notice the trailing dot `.`)

```shell
pip install -v .
```

Download the model file according to the link below.

```
Link (链接): https://pan.baidu.com/s/1HtRb_8mCSFeNtg9iwBtICg?pwd=vj56 
Extraction code (提取码): vj56
```

### 3. Running OASim

Running OASim is straightforward:
```bash
bash scripts/run_oasim.sh
```
**Note:** Modify `--resume_dir` parameter in `run_oasim.sh` to the directory of the model file you downloaded.

#### 3D Preview:

First, move the mouse to the 3D preview interface in the upper right part and click the right mouse button. 
Then preview the 3D implicit reconstruction results through the keyboard. The keyboard control method is as follows.

| Extrinsic Params | Keyboard_input | Extrinsic Params | Keyboard_input |
  | :--------------: | :------------: | :--------------: | :------------: |
  |    +x degree     |       q        |    -x degree     |       a        |
  |    +y degree     |       w        |    -y degree     |       s        |
  |    +z degree     |       e        |    -z degree     |       d        |
  |     +x trans     |       r        |     -x trans     |       f        |
  |     +y trans     |       t        |     -y trans     |       g        |
  |     +z trans     |       y        |     -z trans     |       h        |

#### Traffic flow editor:
First, click the `Choose Start Lane` button in the middle interface, and click on any lane on the HD map to select the starting lane. Then click the `Choose Arrival Lane` button and click on any lane on the HD map to select the destination lane. Finally, click `Confirm` button to generate data.

## 🔖 Citation
If you find our paper and codes useful, please kindly cite us via:

```bibtex
@misc{yan2024oasim,
      title={OASim: an Open and Adaptive Simulator based on Neural Rendering for Autonomous Driving}, 
      author={Guohang Yan and Jiahao Pi and Jianfei Guo and Zhaotong Luo and Min Dou and Nianchen Deng and Qiusheng Huang and Daocheng Fu and Licheng Wen and Pinlong Cai and Xing Gao and Xinyu Cai and Bo Zhang and Xuemeng Yang and Yeqi Bai and Hongbin Zhou and Botian Shi},
      year={2024},
      eprint={2402.03830},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}
```

## 📝 License
OASim is released under the Apache 2.0 license.

## Contact
If you have questions about this repo, please contact Yan Guohang (`yanguohang@pjlab.org.cn`).

OASim

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README