Enhancing 5G Small Cell Selection: A Neural Network and IoV-Based Approach

Area of Science:

• Wireless communication networks
• Telecommunications engineering
• Machine learning applications

Background:

• Ultra-dense networks (UDNs) are crucial for 5G capacity enhancement, but cell selection presents NP-hard computational challenges.
• Internet of Vehicles (IoV) integrates vehicles, infrastructure, and networks to improve transportation systems.
• Efficient cell selection is vital for seamless connectivity in high-density mobile environments.

Purpose of the Study:

• To propose a machine learning and IoV-based cell selection scheme (ANN-CS) for 5G UDNs.
• To optimize cell selection by maximizing dwell time for connected vehicles.
• To reduce computational complexity and improve handover performance.

Main Methods:

• Developed an Artificial Neural Network Cell Selection (ANN-CS) scheme leveraging IoV technology.
• Trained a feed-forward back-propagation ANN (FFBP-ANN) using real-world vehicle and base station data from Los Angeles.
• Focused on selecting small cells with the longest predicted dwell times based on vehicle movement patterns.

Main Results:

• The trained ANN model demonstrated high accuracy with a minimal error rate.
• The ANN-CS scheme reduced the handover rate by up to 33.33%.
• Dwell time was increased by up to 15.47%, minimizing unsuccessful and unnecessary handovers and enhancing downlink throughput.

Conclusions:

• The proposed ANN-CS scheme effectively addresses cell selection challenges in 5G UDNs.
• Integrating machine learning with IoV significantly improves network performance for vehicles.
• The method offers a practical solution for enhancing mobile connectivity and user experience in dense network environments.