Automatic Traffic Scenario Conversion from OpenSCENARIO to CommonRoad

Yuanfei Lin,Michael Ratzel,Matthias Althoff

DOI: https://doi.org/10.1109/ITSC57777.2023.10422422

2023-07-20

Abstract:Scenarios are a crucial element for developing, testing, and verifying autonomous driving systems. However, open-source scenarios are often formulated using different terminologies. This limits their usage across different applications as many scenario representation formats are not directly compatible with each other. To address this problem, we present the first open-source converter from the OpenSCENARIO format to the CommonRoad format, which are two of the most popular scenario formats used in autonomous driving. Our converter employs a simulation tool to execute the dynamic elements defined by OpenSCENARIO. The converter is available at <a class="link-external link-http" href="http://commonroad.in.tum.de" rel="external noopener nofollow">this http URL</a> and we demonstrate its usefulness by converting publicly available scenarios in the OpenSCENARIO format and evaluating them using CommonRoad tools.

Robotics,Formal Languages and Automata Theory,Software Engineering

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the compatibility issue between different autonomous driving scenario description formats, especially the conversion between two widely - used scenario formats, OpenSCENARIO and CommonRoad. Since open - source scenarios are usually defined using different terms, this restricts their use in different applications. To solve this problem, the author has developed an open - source converter from the OpenSCENARIO format to the CommonRoad format. This converter promotes the interoperability and reusability between scenarios of different formats by simulating the dynamic elements defined in OpenSCENARIO through simulation tools and converting the results into the CommonRoad format.

### Main contributions of the paper

1. **The first publicly available converter**: Developed the first open - source converter from OpenSCENARIO to CommonRoad, solving the compatibility problem between the two formats.
2. **Simulation of dynamic elements**: Utilize simulation tools (such as esmini) to execute the dynamic elements defined in OpenSCENARIO, ensuring that the converted scenario can accurately reflect the behavior of the original scenario.
3. **Wide applicability**: This converter is applicable not only to academic research but also to industrial applications, promoting the development, testing, and verification of autonomous driving systems.
4. **Performance evaluation**: Verified the effectiveness and practicality of the converter by converting and evaluating 54 publicly available OpenSCENARIO scenarios.

### Technical implementation

- **Map conversion**: Create a lanelet network based on OpenDRIVE files.
- **Dynamic obstacle trajectory generation**: Obtain the trajectories of all dynamic obstacles through simulation tools.
- **Planning problem construction**: Construct the planning problem in the CommonRoad scenario according to the simulation results.
- **Time - step adjustment**: In order to match the time - step required by CommonRoad, it may be necessary to resample the trajectory.

### Experimental results

- **Conversion success rate**: All test scenarios were successfully converted, with an average conversion time of 31.74 seconds.
- **Scenario evaluation**: Evaluated the converted scenarios through CommonRoad tools in terms of collision detection, motion planning, and safety verification, proving the effectiveness of the converter.

In conclusion, this paper solves the compatibility problem between different scenario formats by developing a converter from OpenSCENARIO to CommonRoad, providing strong support for the development and testing of autonomous driving systems.