Abstract
Road geometry estimation is an essential step in advanced driver assistance systems (ADAS) where an ego vehicle builds a local map of the road ahead using its onboard sensors (Camera, Radar etc.). In the near future, vehicles are also expected to be equipped with dedicated short-range communications (DSRC) transceivers, which would enable them to communicate with nearby vehicles, road-side devices, etc. We propose a novel design for road geometry estimation by fusing on-board sensor (camera and radar) data with data received from remote vehicles (RVs) via DSRC. We propose two measurement models which use standard vehicle-to-vehicle (V2V) messages in road geometry estimation. Our Unscented Kalman Filter (UKF)-based methods fuse measurements by onboard sensors with information from V2V messages to produce a long-range estimate with more than 6.5x the accuracy of current state-of-the-art camera-radar fusion methods.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have