Reinforcement learning based resource allocation scheme for vehicular communication in 5G networks for smart cities

Area of Science:

• Telecommunications Engineering
• Artificial Intelligence
• Smart City Technologies

Background:

• Connected vehicles and infrastructure are rapidly increasing, driving demand for energy-efficient communication.
• Vehicle-to-Everything (V2X) communication is crucial for smart city development, requiring optimized resource management.
• Existing 5G V2X networks face challenges in balancing energy efficiency (EE) with high-quality, low-latency communication.

Purpose of the Study:

• To develop an innovative reinforcement learning (RL)-based method for dynamic resource allocation in 5G V2X networks.
• To enhance energy efficiency (EE) and minimize power consumption in connected vehicle systems.
• To ensure optimal resource utilization while maintaining high-quality service and low-latency communication.

Main Methods:

• Implemented a reinforcement learning (RL) framework for dynamic resource allocation in 5G V2X networks.
• Utilized Q-learning to dynamically adjust transmission power and spectrum allocation in real-time.
• Incorporated factors such as Doppler shift, user mobility, and traffic conditions into the RL model.

Main Results:

• Demonstrated a substantial decrease in power consumption within urban vehicular scenarios.
• Achieved significant improvements in overall network energy efficiency (EE).
• Validated the system's ability to adapt to fluctuating traffic patterns and network demands.

Conclusions:

• The proposed RL-based dynamic resource allocation method offers a sustainable solution for smart mobility.
• The framework effectively enhances energy efficiency and reduces power consumption in 5G V2X networks.
• This advancement promotes greener, more reliable, and energy-efficient urban transportation systems.