Vehicles Selection Algorithm for Cooperative Localization Based on Stochastic Geometry in Internet of Vehicle Systems

Abstract

In the face of the high density of vehicle distribution in urban areas, the self-organized network formed between vehicles has suffered serious communication interference, which will lead to the interruption of communication between vehicles and the inability of collaborative positioning. However, this issue has received little attention, and traditional cooperative positioning methods are no longer suitable for the network of high-density vehicles. In order to solve the cooperative localization issue in high-density vehicle networks, This paper proposes a region-constrained vehicle cooperative localization algorithm based on stochastic geometry. By setting restricted areas of vehicles, the communication capability and network capacity between vehicles are effectively improved. Then, to take full advantage of the density of the Ultra-high density vehicles network, the Geometric Dilution of Precision (GDoP) metric is used to dynamically introduce cooperative vehicles. By dynamically selecting vehicles, the value of GDoP decreases significantly with the increase of vehicle density. Finally, experimental simulations show that the network capacity in the vehicle system is improved by about 71% at the maximum interruption probability when using vehicle selection with restricted areas. Furthermore, the positioning performance of vehicles improves continuously with the increasing density of the cooperative vehicle network.