MapLite 2.0: Online HD Map Inference Using a Prior SD Map

Deploying fully autonomous vehicles has been a subject of intense research in both industry and academia. However, the majority these efforts have relied heavily on High Definition (HD) prior maps. These are necessary to provide planning control modules rich model operating environment. While this approach shown success, it drastically limits scale scope deployments as creating maintaining HD maps for very large areas can be prohibitive. In work, we present new method building map online by starting with Standard (SD) such navigational road map, incorporating onboard sensors infer local map. We evaluate our extensively 100 sequences real-world vehicle data demonstrate that highly structured map-like world accurately using only SD sensors.