FURI | Spring 2022

Autonomous Driving World Map Auto-Generation for CARLA Simulator Using OpenStreetMap

Nowadays, Autonomous Driving Systems (ADS) is a fast-growing field of research. High-fidelity simulation environments ease the assessment of Autonomous Vehicles (AVs) in different safety-critical domains. However, the currently available simulation environments suffer from the lack of diverse simulation scenarios, in particular, map scenarios. To address this issue, the simulator map auto-generation method provides an extensive set of scenarios for AV simulation by enabling the transfer of real-world maps into simulation environments. This research enables more realistic simulations in order to evaluate the performance of AVs in a wide range of safety-critical scenarios using diverse maps of the real world.