Cooperative localization of vehicles without inter-vehicle measurements

Abstract

While cooperation among vehicles can improve localization, standard communication technologies (e.g., 802.11p) cannot provide reliable range or angle measurements. To allow cooperation without explicit inter-vehicle measurements, we propose a cooperative localization method whereby vehicles track mobile features in the environment and use associations of features among vehicles to improve the vehicles' localization accuracy. The proposed algorithm, which scales linearly in the number of vehicles and quadratically in the number of tracked features, shows superior localization performance compared to a non-cooperative approach.