Identifying Value in Crowdsourced Wireless Signal Measurements

A Weighted-Range Classification Model for Localizing Cell using Crowdsource Data

“…Some of the available projects are seen from databases such as OpenCellID, OpenBMap, and OpenSignal. These databases offer a vast amount of dataset for research use as carried out by [2,14] and many more. The studies done involved datasets from various countries and offer insight into how its radio mapping look like while contributing erroneous localization information for future researchers to work on.…”

“…These databases contain datasets gathered from crowdsourcing for both mobile and WiFi. Works by [2,14] Li [14] produced their findings using open cellular databases as well. A study conducted to compare the investments of MNO in the provisioning of mobile coverage in the city against its rural counterpart against income-level indicated that the latter are often excluded or received less attention.…”

“…However, the research proved the worth of crowdsourcing in gathering data from across several operators instead of traditional methods where the focus is usually on a single operator. Li [14] utilized public crowdsourcing database to perform localization prediction also met with poor results and very high variances. However, it is noteworthy to say that the analysis was across several countries and could not have been accomplished without the help of crowdsourcing.…”

“…However, studies do point to weighted centroid as being stable overall [14]. These variances affect the accuracy of the estimated base station localization.…”

“…Some of the available public databases used by researchers are OpenCellID, OpenBMap OpenMobileNetwork and OpenSignal [2].These large databases are loaded with datasets gathered by crowdsourcing approach. Studies by [2,14] indicate that inaccuracies exist within the dataset when estimation algorithms are applied onto it. Some of the known approaches studied are centroid-based approach, weighted centroid based approach, minimum enclosing circle, grid based approach, coverage area estimation (model based), and highest RSS (Radio Signal Strength).…”