Dynamic Generation of a Global Environmental Model for Connected Vehicles

This study explores the enhancement of autonomy in connected vehicles through improved cooperative mapping techniques. By dividing large areas into smaller sections for independent exploration, vehicles can generate precise local maps using integrated sensors such as LIDAR, cameras, and GNSS. The focus is on urban and outdoor environments, utilizing Vehicle-to-Vehicle (V2V) communication rather than Vehicle-to-Everything (V2X) or Vehicle-to-Infrastructure (V2I). Two main map merging methods are evaluated: Direct Map Merging (DMM) and Indirect Map Merging (IMM). DMM offers real-time capabilities and simplicity, while IMM provides improved accuracy and flexibility but is computationally intensive. The study employs the CARLA simulator to validate the proposed cooperative 3D mapping system, utilizing vehicles equipped with advanced sensors. Performance metrics such as fitness score, RMSE, and C2C distance are used to evaluate the effectiveness of the system in different scenarios