Date of Award

11-4-2024

Document Type

Thesis

School

School of Computing

Programme

Ph.D.-Doctoral of Philosophy

First Advisor

Dr.K.R.Manjula

Keywords

WiFi, Indoor Positioning, Dimensionality Reduction, Deep Learning, Accuracy

Abstract

The last few years have seen an increase in interest in indoor positioning and localization as potential research and development areas. WiFi is a strong substitute that supports positioning based on indoor floor plans. In this thesis, the Principal Featured - Kohonen Deep Structure (PF-KDS) model is developed to position WiFi devices more accurately and efficiently for indoor floor planning. Initially, spatial data analysis is conducted using the Principal Feature Enhanced Auto-Encoder algorithm, extracting principal features for dimensionality reduction.

Following this, the Kohonen Self- Organizing Deep Structured Learning technique is devised for precise position estimation by considering a new path loss model incorporating wall influences on Received Signal Strength Indication (RSSI), thereby enhancing device positioning accuracy. The second aspect introduces the Gaussian Distributive Feature Embedding-based Deep Recurrent Perceptive Neural Learning (GDFE-DRPNL) for improved position estimation with reduced errors.

GDFE-DRPNL selects primary features using Gaussian Distributive Feature Embedding and employs Deep Recurrent Multilayer Perceptive Neural Learning with a Deming Regressive Trilateral Positioning Model to compute device positions, resulting in enhanced accuracy. The third research introduces the Linear Geometric Projective Convolutional Deep Belief Network (LGPCDBN) to enhance position estimation accuracy with minimal errors. Dimensionality reduction is initiated through Linear Helmert–Wolf blocked Sammon projection, followed by Geometric Levenberg–Marquardt Convolutional Deep Belief Network for precise device positioning using geometric triangulation methods.

Additionally, the Enhanced River-Formation Dynamics for Multi-node Routing Protocol (ERFD-MRP) is proposed for efficient sensor data collection in Wireless Sensor Networks (WSN) based on River Formation-based Dynamics, ensuring energy-efficient multi-hop routing. Performance evaluations of the proposed methods are conducted using Java language implementation and Indoor Positioning and Indoor Navigation (IPIN) 2016 competition dataset, demonstrating LGPCDBN's superior positional accuracy and ERFD-MRP's higher throughput, packet delivery ratio, lower delay, and energy consumption compared to existing approaches.

Download

Share

COinS