Mobility Management for Cellular-Connected UAVs: A Learning-Based Approach
The pervasiveness of the wireless cellular network can be a key enabler for the deployment of autonomous unmanned aerial vehicles (UAVs) in beyond visual line of sight scenarios without human control. However, traditional cellular networks are optimized for ground user equipment (GUE) such as smartp...
Gespeichert in:
Hauptverfasser: | , , |
---|---|
Format: | Artikel |
Sprache: | eng |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | The pervasiveness of the wireless cellular network can be a key enabler for
the deployment of autonomous unmanned aerial vehicles (UAVs) in beyond visual
line of sight scenarios without human control. However, traditional cellular
networks are optimized for ground user equipment (GUE) such as smartphones
which makes providing connectivity to flying UAVs very challenging. Moreover,
ensuring better connectivity to a moving cellular-connected UAV is notoriously
difficult due to the complex air-to-ground path loss model. In this paper, a
novel mechanism is proposed to ensure robust wireless connectivity and mobility
support for cellular-connected UAVs by tuning the downtilt (DT) angles of all
the GBSs. By leveraging tools from reinforcement learning (RL), DT angles are
dynamically adjusted by using a model-free RL algorithm. The goal is to provide
efficient mobility support in the sky by maximizing the received signal quality
at the UAV while also maintaining good throughput performance of the ground
users. Simulation results show that the proposed RL-based mobility management
(MM) technique can reduce the number of handovers while maintaining the
performance goals, compared to the baseline MM scheme in which the network
always keeps the DT angle fixed. |
---|---|
DOI: | 10.48550/arxiv.2002.01546 |