Influence of Digital Terrain Data Accuracy on Diffraction Prediction in NLOS Wireless Backhauls

Lack of Internet access in most rural and hard-to-reach areas has become a challenge worldwide. Cost-effective non-line-of-sight (NLOS) wireless backhauls via diffraction can help these communities connect to the Internet. Accurate prediction and modelling of the diffraction is necessary for the planning and design of such wireless systems. In this work, we conduct a comparative analysis to examine the influence of the accuracy of various digital elevation models (DEMs) on the predicted diffraction loss for NLOS wireless backhaul through a single blocking obstacle at an unlicensed 5.8 GHz band. The diffraction loss is predicted using the following terrain-based diffraction models: the irregular terrain model (ITM), the knife-edge (KE) diffraction model, and the rounded-obstacle (RO) diffraction model. We have used various freely available DEMs as the main input for the propagation models. Then, we examined the elevation accuracy of these DEMs compared to reference data obtained from light detection and ranging (LiDAR)-based elevation dataset. Measurement data of diffraction loss for various rural sites are used to verify the analytical results. The results show that some DEMs may inaccurately estimate the terrain elevation and obstacle height for the path profile between the transmitter and receiver. This highly impacts the accuracy of the predicted diffraction loss depending on their vertical accuracy and horizontal resolution, particularly in the case of shallow diffraction angles. Thus, it is necessary to evaluate the elevation data accuracy before applying the terrain-based radiofrequency (RF) propagation models. IEEE