Relative Localization of Ground Vehicles Using Non-Terrestrial Networks

Most applications of vehicular networks, such as automatic drive, are based on accurate relative position and orientation information of neighboring vehicles. Due to the geospatial restriction, ground cellular networks are not always available thus non-terrestrial networks such as unmanned aerial vehicles (UAV) and satellites can be used to support the localization of the vehicles. In this paper, we propose an UAV-aided relative localization scheme for ground vehicles. Each node makes time delay and angle measurements with other nodes, and receives GPS signals. The Fisher information matrix (FIM) is derived for the location, orientation and clock bias parameters. The FIM consists of two parts corresponding to the time delay and angle measurements, respectively. Both the absolute and relative localization are studied, and the Cramér-Rao lower bounds for absolute and relative position errors are derived using the information inequality and subspace projection. The simulation results show that the scheme can achieve half-meter relative position accuracy and one-degree orientation accuracy.