Optimizing Wireless Connectivity: A Deep Neural Network-Based Handover Approach for Hybrid LiFi and WiFi Networks.

Area of Science:

• Computer Science
• Electrical Engineering
• Telecommunications

Background:

• Hybrid LiFi and WiFi networks (HLWNet) combine Light Fidelity (LiFi) and Wireless Fidelity (WiFi) for enhanced wireless data transmission.
• Handover decision-making in HLWNet is complex due to LiFi's line-of-sight requirements, differing from traditional heterogeneous networks.

Purpose of the Study:

• To address the intricate handover decision-making challenges in Hybrid LiFi and WiFi networks.
• To propose a novel handover method for HLWNet utilizing deep neural networks (DNNs).

Main Methods:

• The research frames the handover problem as a binary classification task.
• A handover scheme employing both Artificial Neural Networks (ANN) and Deep Neural Networks (DNN) was developed.
• Input factors included user mobility and channel quality for informed handover decisions.

Main Results:

• The DNN-based handover approach achieved over 95% accuracy after training.
• Compared to ANN, the proposed method increased user throughput by 18.58%–38.5%.
• The new method reduced handover rates by 55.21%–67.15% and showed robustness to mobility and channel variations.

Conclusions:

• The proposed DNN-based handover method significantly enhances HLWNet performance.
• The approach offers substantial improvements in user throughput and handover rate reduction.
• This method provides a robust solution for handover management in hybrid wireless networks.