A Novel Low-Cost Method for Characterization of Mobile Propagation Channels with Consumer Devices

Abstract

The latest advancements in mobile device technology are putting ever-higher demands for throughput of wireless networks. This is threatening to outpace the ability of service providers to deploy the necessary infrastructure. Fifth-Generation Mobile Network (5G) technology is experiencing rapid adoption as part of the effort to meet demand, and along with it researchers are continuously seeking new metrics and models for use in predicting the limits of current and future network infrastructure. To succeed, it is key that they have access to methods for simple, effective analysis of the wireless propagation channel in any given location. The typical laboratory test environment lacks the unpredictability and uniqueness of real-world conditions. Additionally, it utilizes equipment whose specifications are often far removed from devices that are actually intended to operate on the mobile network, such as smartphones themselves.

This work focuses on the nature of contemporary path loss models and their ability to accurately predict signal levels, seeking to validate their use against observed path loss behavior in outdoor line-of-sight (LOS) scenarios, where the number of active devices can vary significantly over short periods of time. These conditions are typical of public spaces such as parks and city streets where a large number of users may all simultaneously be accessing high-throughput services. To test their validity, statistics are provided for sets of data collected on foot in public spaces using a novel software utility developed expressly for this purpose. The models we use for comparing against our measured results include both experiential models that are built on other data sets, along with stastically-based, large-scale path loss models. These are compared as a function of distance from the base station (BS), and any unique characteristics of the local network are considered. Finally, a combination of environmental imagery, coverage maps with signal strength overlays, and the aforementioned model comparison are used to estimate the signal source and predict performance in nearby areas.