Measuring And Predicting Cellular Network Quality On Trains

The number of commuters and job-related travelers on public transport is constantly increasing. Often, these spend their time traveling working or consuming news. Consequently demands on mobile network access are rising. Cellular networks, in particular as experienced from within trains, often lack the desired performance. The underlying study exemplary analyses the cellular network quality on a popular commuting route, from which a location dependent quality metric of the cellular network is derived. Based on this it is possible to predict the usability of the network for given service classes (e.g. texting, web browsing, video streaming) for given locations, and also over the duration of the journey. The derived model shows an error of 6.72%, predicting the correct service class in 85% of the locations. Hence, it becomes possible to guide the user in his network utilization by giving feedback on the current state of the network and predicting changes. Thus, the user experience can be improved while simultaneously saving battery power on the smartphone by avoiding or delaying network requests during insufficient network availability.