Resource Allocation using Artificial Intelligence for Vehicle-to-Everything (V2X) Communications on Licensed and Unlicensed Spectrum

This paper presents a Deep Q-Network (DQN) based scheme which adjusts the Transmission Opportunity (TXOP) time for Vehicle-to-Everything (V2X) communications in unlicensed bands. Our scheme adjusts the TXOP duration according to different traffic scenarios and communication requirements to efficiently utilize the unlicensed spectrum. Simulation results demonstrate the effectiveness of our scheme to flexibly allocate unlicensed spectrum resource in different traffic environments and balance the throughput and fairness for V2X and Wi-Fi users.