Abstract
p.p1 {margin: 0.0px 0.0px 0.0px 0.0px; font: 12.0px ‘Times New Roman’;
min-height: 15.0px} p.p2 {margin: 0.0px 0.0px 0.0px 0.0px; font: 9.0px
‘Times New Roman’} span.s1 {font: 12.0px ‘Times New Roman’}
Mobility is a key aspect in current cellular networks, allowing
users to access the provided services almost anywhere. When a user
transitions from a base station’s coverage area to another cell being
serviced by another station, a handoff process takes place, where
resources are released in the first base station, and allocated in the
second for the purpose of servicing the user. Predicting the future
location of a cell phone user allows the handoff process to be
optimized. This optimization allows for a better utilization of the
available resources, regarding bot the transmitted power and the
frequency allocation, resulting in less amount of wasted power in
unwanted directions and the possibility of reusing frequencies in a
single base station. To achieve this goal, Deep Learning techniques are
proposed, which have proven to be efficient tools for predicting and
detecting patterns. The purpose of this paper is to give an overview of
the state of the art in Deep Learning techniques for making
spatio-temporal predictions, which could be used to optimize the handoff
process in cellular systems.