An Experimental Evaluation of Radio Models for Localizing Fixed-Wing UAVs in Rural Environments

Abstract

The emergence of unmanned aerial vehicles (UAVs) has spurred research on radio air-to-ground (AG) channels. The research shows the AG channel can exhibit higher rates of change and more 3D effects than a terrestrial channel. A received-power model predicts the power that makes it from the transmitter (Tx) to the receiver (Rx) through the intervening environment. Received-power models are key to many RF localization schemes. However, many of the received-power models used for UAV localization are 1D or 2D; 3D effects such as ground reflection and polarization are oversimplified or ignored entirely. In this paper, we perform a comparative experimental analysis of various received-power models. The more sophisticated models account for range, Tx directionality, Rx directionality, polarization, and ground reflection. The models are applied to an AG link realized by a fixed-wing UAV in an open-field environment. Our results show that, on average, a 50% reduction in localization error can be achieved by including the polarization factor in the received-power model.