ControlMap: Controllable High-Definition Map Generation for Traffic Scenario Simulation
Researchers have developed ControlMap, a novel pipeline for generating high-definition maps for autonomous driving simulations. This data-driven approach utilizes latent diffusion models and ControlNet for spatial conditioning, allowing for fine-grained control over road topologies. The system also supports adjustable conditioning strength and city-level style transfer, producing realistic maps that adhere to specified road layouts and city details. AI
IMPACT Enables more diverse and targeted scenario generation for autonomous driving simulations, potentially accelerating validation.