Title: Energy efficient deployment of multiple UAV mounted base stations: a machine learning-based approach

Authors: Dilip Mandloi; Rohit Sharma; Rajeev Arya

Addresses: Department of ECE, National Institute of Technology Patna, Patna, Bihar, 800005, India ' Department of ECE, National Institute of Technology Patna, Patna, Bihar, 800005, India ' Department of ECE, National Institute of Technology Patna, Patna, Bihar, 800005, India

Abstract: This paper presents an energy efficient approach for the deployment of multiple 5G enabled unmanned aerial vehicle-mounted base stations (UmBSs) in the area where existing network infrastructure has got demolished due to a natural disaster. In our proposed approach, K-means, a machine learning-based technique, is used to position UmBSs based upon which their transmit power is optimised by solving a convex optimisation problem under the constraints of maximum transmit power of UmBS, maximum height of UmBS, and channel capacity. To ensure the QoS requirements, the association of UEs and UmBS is subjected to the constraint of minimum signal to interference plus noise ratio (SINR) threshold. The effectiveness of the proposed approach is demonstrated in terms of average SINR, average channel capacity, and the number of users served. The performance of the proposed approach is compared with the two baseline approaches through MATLAB simulations.

Keywords: 5G enabled UmBS; machine learning; transmit power optimisation; SINR-based user association.

DOI: 10.1504/IJUWBCS.2022.126775

International Journal of Ultra Wideband Communications and Systems, 2022 Vol.5 No.3, pp.126 - 135

Received: 05 Jun 2021
Accepted: 03 Dec 2021
Published online: 07 Nov 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article