A Multi-model Framework for Tether-based Drone Localization

This paper presents a multi-model localization framework for tethered drones. The framework is composed of three independent localization strategies, each one relying in a different model. The first strategy uses simple trigonometric relations assuming that the tether is taut; the second method relies on a set of catenary equations for the case when there is slack on the tether; the third estimator is a neural network-based predictor that can cover different tether shapes. Multi-layer perceptron networks previously trained with a dataset comprised of the tether variables (i.e., length, tether angles on the drone and on the ground) as inputs are used to select which model instantaneously provides the best results. Those networks are also able to identify situations where tether localization is not possible, thus rejecting all estimates. We evaluate our methodology using an in-house built tethering system, which is also described in this paper. Our experimental results have shown that the proposed localization framework consistently selects good solutions from the three estimators and rejects them when the input tether variables suggest bad estimation results.