| Abstract |
This paper proposes a novel smart communication scheme that can learn and adapt to different impairments and dynamics of vehicular wireless communications. The proposed approach aims to improve Vehicular Ad-hoc Network (VANET) Quality of Service (QoS) through Signal-to-Interference-plus-Noise Ratio (SINR) and outage rate management. The paper contributes by proposing ways to improve SINR and outage rates in VANETs by combining classical control techniques, such as Proportional-Integral-Derivative (PID) controllers, and Reinforcement Learning (RL) capabilities integrated with interference management techniques, such as digital beamforming. A sophisticated simulation platform was used to replicate the scenarios of a roundabout VANET and assess the effectiveness of communication methods and control strategies under fair conditions. In the simulated scenarios with a fixed receiver in the center of the roundabout, RL control methods outperform the PID tuned by the classical Ziegler-Nichols (ZN) tuning technique, ensuring stable SINR values around 20-22 dB when the transmitter and interferer were fixed at roundabout entrances. Furthermore, in scenarios with the transmitter or interferer moving around the roundabout, digital beamforming methods integrated with RL-based controllers showed significant improvements, with outage rates of different methods being reduced by 5-54% compared with non-beamforming or purely beamforming approach, demonstrating the effectiveness of implemented control strategies and highlighting the potential of RL-based techniques in improving VANET communication management. This proposed approach demonstrates suitability for VANET applications, suggesting avenues for further research and development in the field for integration with advanced infrastructures like 5G and 6G for smart cities. © 2013 IEEE. |
