Distributed 3D Relative Localization of UAVs

Abstract

High-accuracy and low-latency localization capability is essential for the internet of things (IoT), such as unmanned aerial vehicles (UAVs). For applications of UAVs such as formation control, the relative positions of the UAVs are more pertinent than their absolute positions. In this paper, we establish a framework for the design and analysis of 3D UAV relative localization. In particular, we develop a distributed algorithm for relative localization networks, where the local geometries are first obtained by a weighted semi-definite programming algorithm and then merged into a global geometry based on the statistical information. Moreover, the relative position estimation is proved to be a constrained optimization where the constraints can eliminate the effect of transformation. The Cramér-Rao lower bound (CRLB) is derived to evaluate the relative localization accuracy, which is equal to the constrained CRLB. The numerical results demonstrate that the proposed algorithm can approach the CRLB and significantly outperform the conventional multidimensional scaling-based algorithm.