Implementing Deep Learning Techniques in 5G IoT Networks for 3D Indoor Positioning: DELTA (DeEp Learning-Based...

Area of Science:

• Wireless communication
• Machine learning
• Indoor localization

Background:

• Fifth-generation wireless (5G) technology promises low latency and high bandwidth, enabling advanced location-based services.
• Integrating 5G with diverse Internet of Things (IoT) technologies presents challenges for precise indoor localization.
• Existing localization methods struggle with the complexities of 3D positioning in heterogeneous 5G IoT networks.

Purpose of the Study:

• To design and evaluate an experimental 5G testbed integrating C-RAN and IoT networks for enhanced 3D indoor localization.
• To propose and implement the DEep Learning-based co-operaTive Architecture (DELTA) machine learning model for precise 3D localization.
• To demonstrate the superiority of the DELTA model over traditional algorithms in 5G IoT environments.

Main Methods:

• Development of a 5G testbed combining C-RAN and various IoT technologies (Zigbee, Raspberry Pi, Arduino, BLE).
• Implementation of the DELTA machine learning model utilizing a 3D multi-layered fingerprint radiomap.
• Recursive estimation of 2D location followed by 3D location prediction for mobile stations.

Main Results:

• The DELTA model successfully improved both vertical and horizontal localization accuracy in a 5G IoT setting.
• The proposed DELTA model demonstrated superior performance compared to Support Vector Machine (SVM) and K-Nearest Neighbor (KNN) algorithms.
• The 3D localization approach using DELTA is effective for multi-floor navigation in complex indoor environments.

Conclusions:

• The DELTA model offers a significant advancement in precise 3D indoor localization within 5G IoT ecosystems.
• The developed 5G testbed provides a valuable platform for researching and validating advanced localization techniques.
• This research paves the way for enhanced applications in smart factories, vertical farms, and smart hospitals requiring accurate indoor positioning.