Composite Filtering for UWB-Based Localization of Quadrotor UAV With Skewed Measurements and Uncertain Dynamics

Nowadays, the ultrawideband (UWB) has become a popular solution to the indoor localization problem of quadrotor unmanned aerial vehicle (UAV) due to its low power consumption and ease of implementation. Nonetheless, the accuracy of UAV localization is affected by the skew-
$t$
measurement noises of the UWB sensor induced mainly by the nonline-of-sight (NLOS) errors and the multipath effects, and the uncertainties in the UAV dynamics caused by external wind disturbances. In this article, a composite filtering approach is proposed for time difference of arrival (TDOA)-based UWB localization of the quadrotor UAV. The main advantage of the composite filtering approach lies in its capability of dealing with the skew-
$t$
measurement noises and the dynamic uncertainties simultaneously. Specifically, the external winds are estimated and rejected in real time via the disturbance observer (DO). Meanwhile, the posterior state distribution under skew-
$t$
measurement noises is updated in an iterative fashion within the variational Bayes (VB) framework. The effectiveness of the proposed approach is validated through both numerical simulation and experimental tests on a quadrotor UAV.