Title:
An Enhanced Deep Neural Network Framework With Ensemble-Driven Feature Selection For Wifi Fingerprinting Indoor Localization
Author:
Supervisor:
Department:
Electrical & Electronics Engineering
Year:
2026
URL:
Abstract:
The rapid growth of the Internet of Things (IoT) and technological advancements has increased demand for accurate and reliable indoor localization systems. However, WiFi fingerprinting in multi-floor ...
View more
Multimedia Content:
01
Methodology
1. WiFi Fingerprinting System Development
The study adopted RSSI-based WiFi fingerprinting because it leverages existing WiFi infrastructure without requiring expensive specialized hardware. During...
View more
02
Citation Analysis
1. Dominant Research Themes
The literature review revolves around:
• Indoor Localization Systems
• WiFi Fingerprinting
• Machine Learning
• Deep Learning
• Feature Selection
• Explainable AI
...
View more
03
Impact
1. Improved Indoor Localization Accuracy
The framework significantly improved localization performance, reducing localization errors across multiple benchmark datasets. On UTPIndoorLoc, the MAE imp...
View more
04
Benefits to Society
1. Safer Indoor Navigation
Accurate indoor localization can assist navigation within hospitals, airports, shopping malls, universities, and emergency facilities. This improves user convenience and ...
View more
05
Applications
1. Smart Buildings
The framework can be integrated into smart buildings for occupant tracking, facility management, and automation systems.
Benefit:
Improves operational efficiency and building...
View more
06
Infographic
Summary of this Thesis
07
Patents / Products
1. ET-RFECV Feature Selection Framework
A novel ensemble-driven feature selection approach combining Extra Trees, Recursive Feature Elimination, and Cross Validation.
Significance:
Improves AP ...
View more
08
Experts
1. Ts. Dr. Azrina Abd Aziz
Institution: Universiti Teknologi PETRONAS (UTP).
Contribution: Principal supervisor overseeing framework design, feature selection development, and explainable AI integ...
View more