Poster: Network-Based UE Mobility Estimation in Mobile Networks

Abstract

The co-existence of small cells and macro cells is a key feature of 4G and future networks. This heterogeneity, together with the increased mobility of user devices can generate a high handover frequency that could lead to unreasonably high call drop probability or poor user experience. By performing smart mobility management, the network can pro-actively adapt to the user and guarantee seamless and smooth cell transitions. In this work, we introduce an algorithm that takes as input sounding reference signal (SRS) measurements available at the base station (eNodeB in 4G systems) to estimate with a low computational requirement the mobility level of the user and with no modification at the user device/equipment (UE) side. The performance of the algorithm is showcased using realistic data and mobility traces. Results show that the classification of UE speed to three mobility classes can be achieved with accuracy of 87% for low mobility, 93% for medium mobility, and 94% for high mobility, respectively.