Cellular network bandwidth prediction in consumer applications

The quality of experience that uses the application for reproduction in movement depends on bandwidth fluctuation during driving. The popularity of technologies that require the delivery of bandwidth information along a defined trajectory has heightened the need to predict bandwidth and examine its behavior. Knowing achievable bandwidth along a defined trajectory, we can assist and adaptively affect video reproduction (bit rate and video resolution). Many solutions used today for estimating bandwidth directly depend on the sequence of previous measurements. Because of fluctuation and other factors that affect, we cannot rely only on previous measurements. In this paper, we propose predicting communication bandwidth using signal strength parameters and other physical factors that can affect the exactness of achievable prediction using machine learning methods.