Development of an Increased-Accuracy System for Clustering Wireless Multipaths

Date of Publication

8-16-2024

Document Type

Dissertation/Thesis

Degree Name

Doctor of Philosophy in Electronics and Communications Engineering

College

Gokongwei College of Engineering

Department/Unit

Electronics And Communications Engg

Honor/Award

Best Paper Award, “An Improved K-Power Means Technique Using Minkowski Distance Metric and Dimension Weights for Clustering Wireless Multipaths in Indoor Channel Scenarios,” 8th International Conference on Computing and Informatics, 09-10 March 2021, Virtual Conference

Thesis Advisor

LAWRENCE Y. MATERUM

Defense Panel Chair

ARGEL A. BANDALA

Defense Panel Member

RECHEL G. ARCILLA

JOEL P. ILAO

MARNEL S. PERADILLA

ROMANO Q. NEYRA

Abstract (English)

This study proposes a developed wireless multipath clustering technique with increased accuracy. The clustering of multipath components is very significant in wireless communications as most current channel models are cluster-based. Grouping these multipath components accurately means an accurate channel model that leads to a reliable performance wireless network. In this study, the COST 2100 channel model served as the ground truth representing different channel scenarios. Pre-processing techniques were done first to the input data through directional cosine transform and whitening transform before being processed by the four main clustering algorithms: Ant Colony Clustering, KPowerMeans, Kernel Power Density-based Algorithm, and Variational Gaussian Mixture Model. Performance analyses were conducted on these techniques for coming up with the proposed clustering technique. The proposed technique is a KPowerMeans derivative that uses the Minkowski distance in which each propagation dimension was weighted using latent scoring from the principal component analysis of the COST 2100 data. The technique was found to provide higher Jaccard-index accuracy for the indoor COST 2100 propagation scenarios. For the semi-urban propagation scenarios, meager accuracy improvements were achieved because the semi-urban environments contain many scatterers leading to a difficulty in clustering. A user interface was made to view the clustering results and serve as a basis for future comparative work.

Abstract Format

html

Language

English

Upload Full Text

wf_yes

Embargo Period

8-15-2024

This document is currently not available here.

Share

COinS