GarchingSim: An Autonomous Driving Simulator with Photorealistic Scenes and Minimalist Workflow

Liguo Zhou,Yinglei Song,Yichao Gao,Zhou Yu,Michael Sodamin,Hongshen Liu,Liang Ma,Lian Liu,Hao Liu,Yang Liu,Haichuan Li,Guang Chen,Alois Knoll

2024-01-30

Abstract:Conducting real road testing for autonomous driving algorithms can be expensive and sometimes impractical, particularly for small startups and research institutes. Thus, simulation becomes an important method for evaluating these algorithms. However, the availability of free and open-source simulators is limited, and the installation and configuration process can be daunting for beginners and interdisciplinary researchers. We introduce an autonomous driving simulator with photorealistic scenes, meanwhile keeping a user-friendly workflow. The simulator is able to communicate with external algorithms through ROS2 or

Robotics,Artificial Intelligence,Computer Vision and Pattern Recognition,Systems and Control

What problem does this paper attempt to address?

The problem that this paper attempts to solve is that in the field of autonomous driving, the cost of actual road tests is high and difficult to carry out, especially for small start - up companies and research institutions. Therefore, simulation has become an important method for evaluating autonomous driving algorithms. However, the number of existing free and open - source simulators is limited, and the installation and configuration processes can be very difficult for beginners and interdisciplinary researchers. In response to these problems, the author introduces an autonomous driving simulator - GarchingSim, which has photo - realistic scenes and a minimalist workflow.

Specifically, this simulator aims to:
1. **Provide a highly realistic simulation environment**: By using the Unity 3D game engine, GarchingSim can generate highly realistic 3D urban environments, including buildings, roads, vehicles, etc.
2. **Simplify the deployment and use processes**: Ensure that beginners and interdisciplinary researchers can easily get started, emphasizing the "plug - and - play" design concept.
3. **Support multiple sensors**: Including cameras, radars, inertial measurement units (IMU) and global navigation satellite systems (GNSS) to meet the needs of different application scenarios.
4. **Achieve an accurate vehicle dynamics model**: Enhance the realism of vehicle physics.
5. **Be compatible with external software**: Communicate with existing software stacks through ROS2 or Socket.IO to ensure the flexibility and extensibility of the system.
6. **Generate synthetic data**: Used for training and validating machine - learning algorithms, especially in the field of deep learning.
7. **Support multiple applications**: Including imitation learning, reinforcement learning, vehicle - to - vehicle (V2V) and vehicle - to - everything (V2X) communications, etc.

Through these features, GarchingSim aims to provide researchers and developers with an efficient, easy - to - use and powerful autonomous driving simulation platform.