Polarimetric Radar for Automotive Self-Localization

Abstract

In order to enable highly automated driving, high demands are placed on the sensors perceiving the environment around the vehicle. Today's radar sensors have advantages over other sensor types because they are robust in adverse weather conditions and enable an instantaneous velocity estimation. Newly available automotive radar sensors with polarimetric capabilities provide even richer information on the perceived objects in the environment including, for example, shapes and orientations. This work shows how the additional information can be incorporated into gridmaps that are especially useful for vehicle self-localization tasks. Experiments have been carried out to demonstrate real world results.